associations.Rmd

Summary of analysis

The various rulesets created returned the following trends:

• There are relatively few customers that give a rating of 5, but those that do have the following attributes: male, business trip, economy class, flight not cancelled, price_sensitivity=1
• Female business travelers–while few–also give higher ratings (4-5). They expect no cancellations and low delays. No female business travelers were Blue status in any of the rules; this correlates well with the negative coefficient for Blue status in the linear model.
• Female Blue status customers, traveling for personal reasons in the economy class give the lowest ratings
• Cheapseats airlines with airline_code=WN recieves high ratings from female travelers age (34-46)

Miscellaneous Observations

• Surprisingly loyalty cards don’t seem to influence higher satisfaction ratings. Loyalty cards were hardly present in pairings, and when they were, it was for 0 cards.
• As an example: the top 3 of the 32 rules returned for female business travelers, age 46-58, were for 0 loyalty cards

TODO

• Add the proportion of travelers in the assessment to increase the impact of the assessment. Who cares about the assesment if it accounts for only a small proportion of the population

Unanswered business questions

• What airline recieves the worst ratings from women/men?
• What airline recieves the best ratings from women/men?
• What locations recieve the best/worst ratings?

import data

df <- read.csv("flight_survey_updated.csv", stringsAsFactors = T)

discretize the following values

• cut returns a factor with the levels specified in the array indicated with (start, finish]; start is exclusive, finish is inclusive

df$satisfaction <- cut(df$satisfaction, c(0,1,2,3,4,5))
df$age <- cut(df$age, c(14,34,46,58,86)) 
df$price_sensitivity <- cut(df$price_sensitivity, c(-1,0,1,3,6))
df$year_first_flight <- cut(df$year_first_flight, breaks=3)
df$num_flights <- cut(df$num_flights, c(-1,0,9,18,21,30,101))
df$percent_flight_other_airlines <- cut(df$percent_flight_other_airlines, c(0,4,7,9.2,10,51))
df$num_loyalty_cards <- cut(df$num_loyalty_cards, c(-1,0,1,2,3,7,11,13))
df$airport_shopping <- cut(df$airport_shopping, c(-1,0,27,30,880))
df$airport_dining <- cut(df$airport_dining, c(-1,0,31,61,69,91,896))
df$day_of_month <- cut(df$day_of_month, c(0,1,9,17,24,32))
df$scheduled_departure_hour <- cut(df$scheduled_departure_hour, c(0,1,10,14,18,23))
df$flight_time_in_minutes <- cut(df$flight_time_in_minutes, c(7,61,95,112,142,670))
df$flight_distance <- cut(df$flight_distance, c(0,32,363,630,794,1025,4984))

remove arrival_delay_in_minutes column

df <- df[,-which(colnames(df)=="arrival_delay_in_minutes")]

check for NA values

sapply(df, function(x) sum(is.na(x)))

##                  satisfaction                airline_status 
##                             0                             0 
##                           age                        gender 
##                             0                             0 
##             price_sensitivity             year_first_flight 
##                             0                             0 
##                   num_flights percent_flight_other_airlines 
##                             0                             0 
##                type_of_travel             num_loyalty_cards 
##                             0                             0 
##              airport_shopping                airport_dining 
##                             0                             0 
##                         class                  day_of_month 
##                             0                             0 
##                   flight_date                  airline_code 
##                             0                             0 
##                  airline_name                   origin_city 
##                             0                             0 
##                  origin_state              destination_city 
##                             0                             0 
##             destination_state      scheduled_departure_hour 
##                             0                             0 
##    departure_delay_in_minutes              flight_cancelled 
##                             0                             0 
##        flight_time_in_minutes               flight_distance 
##                             0                             0 
##  arrival_delay_greater_5_mins 
##                             0

change departure_delay_in_minutes to a factor of yes or no if the departure delay is greater than 15 minutes

• change the name of the column to departure_delay_greater_5_mins

df$departure_delay_in_minutes <- ifelse(df$departure_delay_in_minutes > 15,"yes","no")
colnames(df)[which(colnames(df)=="departure_delay_in_minutes")] = "departure_delay_greater_5_mins"
df$departure_delay_greater_5_mins <- factor(df$departure_delay_greater_5_mins)

create sparse matrix from df

dfX <- as(df, "transactions")

inspect the first 2 rows from the matrix

show the item frequency to see which categories are present the most in the matrix

itemFrequencyPlot(dfX, support=.1, cex.names=.7)

Let’s analyze flyers by the ratings groups of 1-5

customer satisfaction rating = 1

ruleset <- apriori(dfX, parameter=list(support=.01, confidence=.5, minlen=2, maxlen=6, maxtime=10),
                   appearance = list(lhs="satisfaction=(0,1]"))

## Apriori
## 
## Parameter specification:
##  confidence minval smax arem  aval originalSupport maxtime support minlen
##         0.5    0.1    1 none FALSE            TRUE      10    0.01      2
##  maxlen target   ext
##       6  rules FALSE
## 
## Algorithmic control:
##  filter tree heap memopt load sort verbose
##     0.1 TRUE TRUE  FALSE TRUE    2    TRUE
## 
## Absolute minimum support count: 1294 
## 
## set item appearances ...[1 item(s)] done [0.00s].
## set transactions ...[896 item(s), 129435 transaction(s)] done [0.42s].
## sorting and recoding items ... [282 item(s)] done [0.05s].
## creating transaction tree ... done [0.10s].
## checking subsets of size 1 2 done [0.01s].
## writing ... [10 rule(s)] done [0.00s].
## creating S4 object  ... done [0.04s].

plot(ruleset)

ruleset_filter <- subset(ruleset, subset=lhs %ain% c("satisfaction=(0,1]") & lift>1)
plot(ruleset_filter)

summary(ruleset_filter)

## set of 8 rules
## 
## rule length distribution (lhs + rhs):sizes
## 2 
## 8 
## 
##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##       2       2       2       2       2       2 
## 
## summary of quality measures:
##     support          confidence          lift           count     
##  Min.   :0.01313   Min.   :0.5688   Min.   :1.000   Min.   :1700  
##  1st Qu.:0.01380   1st Qu.:0.5977   1st Qu.:1.018   1st Qu.:1786  
##  Median :0.01447   Median :0.6265   Median :1.136   Median :1872  
##  Mean   :0.01639   Mean   :0.7099   Mean   :1.278   Mean   :2122  
##  3rd Qu.:0.01941   3rd Qu.:0.8405   3rd Qu.:1.363   3rd Qu.:2512  
##  Max.   :0.02267   Max.   :0.9816   Max.   :2.001   Max.   :2934  
## 
## mining info:
##  data ntransactions support confidence
##   dfX        129435    0.01        0.5

ruleset_filter <- sort(ruleset_filter, decreasing=T, by="count")
inspect(ruleset_filter)

##     lhs                     rhs                                   support confidence     lift count
## [1] {satisfaction=(0,1]} => {flight_cancelled=No}              0.02266775  0.9815992 1.000065  2934
## [2] {satisfaction=(0,1]} => {airline_status=Blue}              0.01999459  0.8658414 1.263959  2588
## [3] {satisfaction=(0,1]} => {class=Economy}                    0.01921428  0.8320509 1.022516  2487
## [4] {satisfaction=(0,1]} => {num_loyalty_cards=(-1,0]}         0.01464828  0.6343259 1.197200  1896
## [5] {satisfaction=(0,1]} => {type_of_travel=Personal}          0.01428516  0.6186015 2.001167  1849
## [6] {satisfaction=(0,1]} => {gender=Female}                    0.01399930  0.6062228 1.073941  1812
## [7] {satisfaction=(0,1]} => {airport_shopping=(-1,0]}          0.01321126  0.5720977 1.006367  1710
## [8] {satisfaction=(0,1]} => {arrival_delay_greater_5_mins=yes} 0.01313401  0.5687521 1.659485  1700

• there are relatively few customers who give a customer satisfaction rating of 1, but those that do have one or more of the following attributes: female, personal travel, arrivel delay > 5 mins, blue status, economy class, 0 loyalty cards, and 0 airport_shopping

customer satisfaction rating = 2

ruleset <- apriori(dfX, parameter=list(support=.1, confidence=.5, minlen=2, maxtime=10),
                   appearance = list(lhs="satisfaction=(1,2]"))

## Apriori
## 
## Parameter specification:
##  confidence minval smax arem  aval originalSupport maxtime support minlen
##         0.5    0.1    1 none FALSE            TRUE      10     0.1      2
##  maxlen target   ext
##      10  rules FALSE
## 
## Algorithmic control:
##  filter tree heap memopt load sort verbose
##     0.1 TRUE TRUE  FALSE TRUE    2    TRUE
## 
## Absolute minimum support count: 12943 
## 
## set item appearances ...[1 item(s)] done [0.00s].
## set transactions ...[896 item(s), 129435 transaction(s)] done [0.42s].
## sorting and recoding items ... [73 item(s)] done [0.04s].
## creating transaction tree ... done [0.09s].
## checking subsets of size 1 2 done [0.02s].
## writing ... [9 rule(s)] done [0.00s].
## creating S4 object  ... done [0.03s].

ruleset_filter <- subset(ruleset, subset=lhs %ain% c("satisfaction=(1,2]") & lift>1)
plot(ruleset_filter)

summary(ruleset_filter)

## set of 6 rules
## 
## rule length distribution (lhs + rhs):sizes
## 2 
## 6 
## 
##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##       2       2       2       2       2       2 
## 
## summary of quality measures:
##     support         confidence          lift           count      
##  Min.   :0.1117   Min.   :0.6150   Min.   :1.013   Min.   :14452  
##  1st Qu.:0.1137   1st Qu.:0.6264   1st Qu.:1.086   1st Qu.:14719  
##  Median :0.1311   Median :0.7220   Median :1.158   Median :16967  
##  Mean   :0.1330   Mean   :0.7325   Mean   :1.383   Mean   :17212  
##  3rd Qu.:0.1485   3rd Qu.:0.8181   3rd Qu.:1.280   3rd Qu.:19224  
##  Max.   :0.1618   Max.   :0.8912   Max.   :2.583   Max.   :20942  
## 
## mining info:
##  data ntransactions support confidence
##   dfX        129435     0.1        0.5

ruleset_filter <- sort(ruleset_filter, decreasing=T, by="count")
inspect(ruleset_filter)

##     lhs                     rhs                        support  
## [1] {satisfaction=(1,2]} => {airline_status=Blue}      0.1617955
## [2] {satisfaction=(1,2]} => {class=Economy}            0.1497122
## [3] {satisfaction=(1,2]} => {type_of_travel=Personal}  0.1449608
## [4] {satisfaction=(1,2]} => {num_loyalty_cards=(-1,0]} 0.1172094
## [5] {satisfaction=(1,2]} => {gender=Female}            0.1125507
## [6] {satisfaction=(1,2]} => {airport_shopping=(-1,0]}  0.1116545
##     confidence lift     count
## [1] 0.8911869  1.300958 20942
## [2] 0.8246308  1.013398 19378
## [3] 0.7984595  2.583005 18763
## [4] 0.6456019  1.218482 15171
## [5] 0.6199413  1.098244 14568
## [6] 0.6150049  1.081844 14452

• Virtually the same result from customers with ratings=1; female, personal travel, blue status, economy class, 0 loyalty cards, and 0 airport_shopping

customer satisfaction rating = 3

ruleset <- apriori(dfX, parameter=list(support=.1, confidence=.5, minlen=2, maxlen=6, maxtime=10),
                   appearance = list(lhs="satisfaction=(2,3]"))

## Apriori
## 
## Parameter specification:
##  confidence minval smax arem  aval originalSupport maxtime support minlen
##         0.5    0.1    1 none FALSE            TRUE      10     0.1      2
##  maxlen target   ext
##       6  rules FALSE
## 
## Algorithmic control:
##  filter tree heap memopt load sort verbose
##     0.1 TRUE TRUE  FALSE TRUE    2    TRUE
## 
## Absolute minimum support count: 12943 
## 
## set item appearances ...[1 item(s)] done [0.00s].
## set transactions ...[896 item(s), 129435 transaction(s)] done [0.45s].
## sorting and recoding items ... [73 item(s)] done [0.04s].
## creating transaction tree ... done [0.08s].
## checking subsets of size 1 2 done [0.04s].
## writing ... [9 rule(s)] done [0.00s].
## creating S4 object  ... done [0.03s].

plot(ruleset)

ruleset_filter <- subset(ruleset, subset=lhs %ain% c("satisfaction=(2,3]") & lift>1)
plot(ruleset_filter)

summary(ruleset_filter)

## set of 7 rules
## 
## rule length distribution (lhs + rhs):sizes
## 2 
## 7 
## 
##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##       2       2       2       2       2       2 
## 
## summary of quality measures:
##     support         confidence          lift           count      
##  Min.   :0.1548   Min.   :0.5434   Min.   :1.005   Min.   :20033  
##  1st Qu.:0.1726   1st Qu.:0.6060   1st Qu.:1.022   1st Qu.:22342  
##  Median :0.1925   Median :0.6760   Median :1.027   Median :24920  
##  Mean   :0.1955   Mean   :0.6865   Mean   :1.047   Mean   :25310  
##  3rd Qu.:0.2217   3rd Qu.:0.7784   3rd Qu.:1.071   3rd Qu.:28696  
##  Max.   :0.2328   Max.   :0.8175   Max.   :1.115   Max.   :30138  
## 
## mining info:
##  data ntransactions support confidence
##   dfX        129435     0.1        0.5

ruleset_filter <- sort(ruleset_filter, decreasing=T, by="count")
inspect(ruleset_filter)

##     lhs                     rhs                                   support confidence     lift count
## [1] {satisfaction=(2,3]} => {class=Economy}                     0.2328427  0.8175012 1.004636 30138
## [2] {satisfaction=(2,3]} => {departure_delay_greater_5_mins=no} 0.2257658  0.7926545 1.017900 29222
## [3] {satisfaction=(2,3]} => {airline_status=Blue}               0.2176382  0.7641187 1.115464 28170
## [4] {satisfaction=(2,3]} => {arrival_delay_greater_5_mins=no}   0.1925291  0.6759616 1.028435 24920
## [5] {satisfaction=(2,3]} => {gender=Female}                     0.1789161  0.6281669 1.112816 23158
## [6] {satisfaction=(2,3]} => {airport_shopping=(-1,0]}           0.1663151  0.5839256 1.027173 21527
## [7] {satisfaction=(2,3]} => {num_loyalty_cards=(-1,0]}          0.1547727  0.5434004 1.025591 20033

• Similar assessment to customers with rating=2 except for one key difference: arrival and departure delays are NOT greater than 5 mins

customer satisfaction rating = 4

ruleset <- apriori(dfX, parameter=list(support=.1, confidence=.5, minlen=2, maxtime=10),
                   appearance = list(lhs="satisfaction=(3,4]"))

## Apriori
## 
## Parameter specification:
##  confidence minval smax arem  aval originalSupport maxtime support minlen
##         0.5    0.1    1 none FALSE            TRUE      10     0.1      2
##  maxlen target   ext
##      10  rules FALSE
## 
## Algorithmic control:
##  filter tree heap memopt load sort verbose
##     0.1 TRUE TRUE  FALSE TRUE    2    TRUE
## 
## Absolute minimum support count: 12943 
## 
## set item appearances ...[1 item(s)] done [0.00s].
## set transactions ...[896 item(s), 129435 transaction(s)] done [0.43s].
## sorting and recoding items ... [73 item(s)] done [0.04s].
## creating transaction tree ... done [0.09s].
## checking subsets of size 1 2 done [0.04s].
## writing ... [9 rule(s)] done [0.00s].
## creating S4 object  ... done [0.03s].

plot(ruleset)

ruleset_filter <- subset(ruleset, subset=lhs %ain% c("satisfaction=(3,4]") & lift>1)
plot(ruleset_filter)

summary(ruleset_filter)

## set of 5 rules
## 
## rule length distribution (lhs + rhs):sizes
## 2 
## 5 
## 
##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##       2       2       2       2       2       2 
## 
## summary of quality measures:
##     support         confidence          lift           count      
##  Min.   :0.2900   Min.   :0.7007   Min.   :1.004   Min.   :37539  
##  1st Qu.:0.3055   1st Qu.:0.7381   1st Qu.:1.034   1st Qu.:39540  
##  Median :0.3415   Median :0.8251   Median :1.060   Median :44202  
##  Mean   :0.3379   Mean   :0.8164   Mean   :1.116   Mean   :43735  
##  3rd Qu.:0.3446   3rd Qu.:0.8325   3rd Qu.:1.123   3rd Qu.:44602  
##  Max.   :0.4079   Max.   :0.9854   Max.   :1.357   Max.   :52792  
## 
## mining info:
##  data ntransactions support confidence
##   dfX        129435     0.1        0.5

ruleset_filter <- sort(ruleset_filter, decreasing=T, by="count")
inspect(ruleset_filter)

##     lhs                     rhs                                   support confidence     lift count
## [1] {satisfaction=(3,4]} => {flight_cancelled=No}               0.4078650  0.9854218 1.003960 52792
## [2] {satisfaction=(3,4]} => {type_of_travel=Business}           0.3445899  0.8325462 1.357443 44602
## [3] {satisfaction=(3,4]} => {departure_delay_greater_5_mins=no} 0.3414996  0.8250798 1.059540 44202
## [4] {satisfaction=(3,4]} => {arrival_delay_greater_5_mins=no}   0.3054815  0.7380584 1.122912 39540
## [5] {satisfaction=(3,4]} => {price_sensitivity=(0,1]}           0.2900220  0.7007074 1.033727 37539

• Customers with one or more of the following attributes (business, departure/arrival delay < 5 min, flight not cancelled, price_sensitivity=1) give ratings=4

customer satisfaction rating = 5

ruleset <- apriori(dfX, parameter=list(support=.01, confidence=.5, minlen=2, maxlen=6, maxtime=10),
                   appearance = list(lhs="satisfaction=(4,5]"))

## Apriori
## 
## Parameter specification:
##  confidence minval smax arem  aval originalSupport maxtime support minlen
##         0.5    0.1    1 none FALSE            TRUE      10    0.01      2
##  maxlen target   ext
##       6  rules FALSE
## 
## Algorithmic control:
##  filter tree heap memopt load sort verbose
##     0.1 TRUE TRUE  FALSE TRUE    2    TRUE
## 
## Absolute minimum support count: 1294 
## 
## set item appearances ...[1 item(s)] done [0.00s].
## set transactions ...[896 item(s), 129435 transaction(s)] done [0.46s].
## sorting and recoding items ... [282 item(s)] done [0.05s].
## creating transaction tree ... done [0.09s].
## checking subsets of size 1 2 done [0.02s].
## writing ... [8 rule(s)] done [0.00s].
## creating S4 object  ... done [0.04s].

ruleset_filter <- subset(ruleset, subset=lhs %ain% c("satisfaction=(4,5]") & lift>1)
plot(ruleset_filter)

summary(ruleset_filter)

## set of 5 rules
## 
## rule length distribution (lhs + rhs):sizes
## 2 
## 5 
## 
##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##       2       2       2       2       2       2 
## 
## summary of quality measures:
##     support          confidence          lift           count      
##  Min.   :0.06416   Min.   :0.6640   Min.   :1.005   Min.   : 8305  
##  1st Qu.:0.07536   1st Qu.:0.7798   1st Qu.:1.016   1st Qu.: 9754  
##  Median :0.07901   Median :0.8176   Median :1.150   Median :10227  
##  Mean   :0.08101   Mean   :0.8383   Mean   :1.243   Mean   :10486  
##  3rd Qu.:0.09017   3rd Qu.:0.9331   3rd Qu.:1.521   3rd Qu.:11671  
##  Max.   :0.09636   Max.   :0.9971   Max.   :1.525   Max.   :12472  
## 
## mining info:
##  data ntransactions support confidence
##   dfX        129435    0.01        0.5

ruleset_filter <- sort(ruleset_filter, decreasing=T, by="lift")
inspect(ruleset_filter)

##     lhs                     rhs                       support   
## [1] {satisfaction=(4,5]} => {gender=Male}             0.06416348
## [2] {satisfaction=(4,5]} => {type_of_travel=Business} 0.09016881
## [3] {satisfaction=(4,5]} => {price_sensitivity=(0,1]} 0.07535829
## [4] {satisfaction=(4,5]} => {flight_cancelled=No}     0.09635724
## [5] {satisfaction=(4,5]} => {class=Economy}           0.07901263
##     confidence lift     count
## [1] 0.6639751  1.524571  8305
## [2] 0.9330828  1.521365 11671
## [3] 0.7798209  1.150440  9754
## [4] 0.9971218  1.015880 12472
## [5] 0.8176367  1.004802 10227

• Customers with one or more of the following attributes give a rating=5: male, business trip, economy class, price sensitivity=1, flight not cancelled

Let’s dig deeper into why females give lower ratings

female_ruleset <- apriori(dfX, parameter=list(support=.01, confidence=.8, minlen=2, maxlen=6, maxtime=15))

## Apriori
## 
## Parameter specification:
##  confidence minval smax arem  aval originalSupport maxtime support minlen
##         0.8    0.1    1 none FALSE            TRUE      15    0.01      2
##  maxlen target   ext
##       6  rules FALSE
## 
## Algorithmic control:
##  filter tree heap memopt load sort verbose
##     0.1 TRUE TRUE  FALSE TRUE    2    TRUE
## 
## Absolute minimum support count: 1294 
## 
## set item appearances ...[0 item(s)] done [0.00s].
## set transactions ...[896 item(s), 129435 transaction(s)] done [0.44s].
## sorting and recoding items ... [282 item(s)] done [0.05s].
## creating transaction tree ... done [0.09s].
## checking subsets of size 1 2 3 4 5 6

## Warning in apriori(dfX, parameter = list(support = 0.01, confidence =
## 0.8, : Mining stopped (maxlen reached). Only patterns up to a length of 6
## returned!

##  done [29.46s].
## writing ... [1158073 rule(s)] done [0.20s].
## creating S4 object  ... done [0.24s].

female_filter <- subset(female_ruleset, subset=lhs %ain% c("gender=Female","satisfaction=(0,1]") & lift>1)
plot(female_filter)

female_filter <- sort(female_filter, decreasing=T, by="count")
summary(female_filter)

## set of 7 rules
## 
## rule length distribution (lhs + rhs):sizes
## 3 4 5 
## 1 4 2 
## 
##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##   3.000   4.000   4.000   4.143   4.500   5.000 
## 
## summary of quality measures:
##     support          confidence          lift           count     
##  Min.   :0.01031   Min.   :0.8153   Min.   :1.001   Min.   :1334  
##  1st Qu.:0.01040   1st Qu.:0.9209   1st Qu.:1.002   1st Qu.:1346  
##  Median :0.01119   Median :0.9260   Median :1.344   Median :1448  
##  Mean   :0.01131   Mean   :0.9250   Mean   :1.200   Mean   :1463  
##  3rd Qu.:0.01195   3rd Qu.:0.9543   3rd Qu.:1.348   3rd Qu.:1546  
##  Max.   :0.01296   Max.   :0.9831   Max.   :1.352   Max.   :1678  
## 
## mining info:
##  data ntransactions support confidence
##   dfX        129435    0.01        0.8

inspect(female_filter)

##     lhs                      rhs                      support confidence     lift count
## [1] {satisfaction=(0,1],                                                               
##      gender=Female}       => {airline_status=Blue} 0.01296404  0.9260486 1.351850  1678
## [2] {satisfaction=(0,1],                                                               
##      gender=Female,                                                                    
##      flight_cancelled=No} => {airline_status=Blue} 0.01270908  0.9262387 1.352127  1645
## [3] {satisfaction=(0,1],                                                               
##      gender=Female,                                                                    
##      class=Economy}       => {flight_cancelled=No} 0.01118708  0.9823609 1.000841  1448
## [4] {satisfaction=(0,1],                                                               
##      gender=Female,                                                                    
##      flight_cancelled=No} => {class=Economy}       0.01118708  0.8153153 1.001950  1448
## [5] {satisfaction=(0,1],                                                               
##      gender=Female,                                                                    
##      class=Economy}       => {airline_status=Blue} 0.01048403  0.9206242 1.343931  1357
## [6] {satisfaction=(0,1],                                                               
##      airline_status=Blue,                                                              
##      gender=Female,                                                                    
##      class=Economy}       => {flight_cancelled=No} 0.01030633  0.9830508 1.001544  1334
## [7] {satisfaction=(0,1],                                                               
##      gender=Female,                                                                    
##      class=Economy,                                                                    
##      flight_cancelled=No} => {airline_status=Blue} 0.01030633  0.9212707 1.344875  1334

• 7 rules (sored decreasing by count, confidence > .9, lift > 1) were generated for female travelers (rating = 1) with the following attributes: flight not cancelled, blue status, economy class

female_ruleset <- apriori(dfX, parameter=list(support=.01, confidence=.8, minlen=2, maxlen=6, maxtime=10))

## Apriori
## 
## Parameter specification:
##  confidence minval smax arem  aval originalSupport maxtime support minlen
##         0.8    0.1    1 none FALSE            TRUE      10    0.01      2
##  maxlen target   ext
##       6  rules FALSE
## 
## Algorithmic control:
##  filter tree heap memopt load sort verbose
##     0.1 TRUE TRUE  FALSE TRUE    2    TRUE
## 
## Absolute minimum support count: 1294 
## 
## set item appearances ...[0 item(s)] done [0.00s].
## set transactions ...[896 item(s), 129435 transaction(s)] done [0.40s].
## sorting and recoding items ... [282 item(s)] done [0.05s].
## creating transaction tree ... done [0.10s].
## checking subsets of size 1 2 3 4 5

## Warning in apriori(dfX, parameter = list(support = 0.01, confidence =
## 0.8, : Mining stopped (time limit reached). Only patterns up to a length of
## 5 returned!

##  done [13.04s].
## writing ... [533271 rule(s)] done [0.08s].
## creating S4 object  ... done [0.12s].

female_filter <- subset(female_ruleset, subset=lhs %ain% c("gender=Female","satisfaction=(1,2]") & lift>8)
plot(female_filter)

## To reduce overplotting, jitter is added! Use jitter = 0 to prevent jitter.

female_filter <- sort(female_filter, decreasing=T, by="count")
summary(female_filter)

## set of 18 rules
## 
## rule length distribution (lhs + rhs):sizes
##  4  5 
##  4 14 
## 
##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##   4.000   5.000   5.000   4.778   5.000   5.000 
## 
## summary of quality measures:
##     support          confidence      lift           count     
##  Min.   :0.01002   Min.   :1    Min.   :8.432   Min.   :1297  
##  1st Qu.:0.01143   1st Qu.:1    1st Qu.:8.432   1st Qu.:1479  
##  Median :0.01169   Median :1    Median :8.432   Median :1513  
##  Mean   :0.01214   Mean   :1    Mean   :8.854   Mean   :1571  
##  3rd Qu.:0.01308   3rd Qu.:1    3rd Qu.:9.382   3rd Qu.:1693  
##  Max.   :0.01458   Max.   :1    Max.   :9.382   Max.   :1887  
## 
## mining info:
##  data ntransactions support confidence
##   dfX        129435    0.01        0.8

inspect(female_filter)

##      lhs                          rhs                         support confidence     lift count
## [1]  {satisfaction=(1,2],                                                                      
##       gender=Female,                                                                           
##       airline_code=EV}         => {airline_name=FlyFast}   0.01457875          1 8.432248  1887
## [2]  {satisfaction=(1,2],                                                                      
##       gender=Female,                                                                           
##       airline_name=FlyFast}    => {airline_code=EV}        0.01457875          1 8.432248  1887
## [3]  {satisfaction=(1,2],                                                                      
##       gender=Female,                                                                           
##       airline_code=EV,                                                                         
##       flight_cancelled=No}     => {airline_name=FlyFast}   0.01386796          1 8.432248  1795
## [4]  {satisfaction=(1,2],                                                                      
##       gender=Female,                                                                           
##       airline_name=FlyFast,                                                                    
##       flight_cancelled=No}     => {airline_code=EV}        0.01386796          1 8.432248  1795
## [5]  {satisfaction=(1,2],                                                                      
##       airline_status=Blue,                                                                     
##       gender=Female,                                                                           
##       airline_code=EV}         => {airline_name=FlyFast}   0.01307992          1 8.432248  1693
## [6]  {satisfaction=(1,2],                                                                      
##       airline_status=Blue,                                                                     
##       gender=Female,                                                                           
##       airline_name=FlyFast}    => {airline_code=EV}        0.01307992          1 8.432248  1693
## [7]  {satisfaction=(1,2],                                                                      
##       gender=Female,                                                                           
##       type_of_travel=Personal,                                                                 
##       airline_code=EV}         => {airline_name=FlyFast}   0.01249276          1 8.432248  1617
## [8]  {satisfaction=(1,2],                                                                      
##       gender=Female,                                                                           
##       type_of_travel=Personal,                                                                 
##       airline_name=FlyFast}    => {airline_code=EV}        0.01249276          1 8.432248  1617
## [9]  {satisfaction=(1,2],                                                                      
##       gender=Female,                                                                           
##       class=Economy,                                                                           
##       airline_code=EV}         => {airline_name=FlyFast}   0.01168926          1 8.432248  1513
## [10] {satisfaction=(1,2],                                                                      
##       gender=Female,                                                                           
##       class=Economy,                                                                           
##       airline_name=FlyFast}    => {airline_code=EV}        0.01168926          1 8.432248  1513
## [11] {satisfaction=(1,2],                                                                      
##       gender=Female,                                                                           
##       airline_name=Northwest}  => {airline_code=OO}        0.01161973          1 9.382067  1504
## [12] {satisfaction=(1,2],                                                                      
##       gender=Female,                                                                           
##       airline_code=OO}         => {airline_name=Northwest} 0.01161973          1 9.382067  1504
## [13] {satisfaction=(1,2],                                                                      
##       gender=Female,                                                                           
##       airline_name=Northwest,                                                                  
##       flight_cancelled=No}     => {airline_code=OO}        0.01142658          1 9.382067  1479
## [14] {satisfaction=(1,2],                                                                      
##       gender=Female,                                                                           
##       airline_code=OO,                                                                         
##       flight_cancelled=No}     => {airline_name=Northwest} 0.01142658          1 9.382067  1479
## [15] {satisfaction=(1,2],                                                                      
##       airline_status=Blue,                                                                     
##       gender=Female,                                                                           
##       airline_name=Northwest}  => {airline_code=OO}        0.01047630          1 9.382067  1356
## [16] {satisfaction=(1,2],                                                                      
##       airline_status=Blue,                                                                     
##       gender=Female,                                                                           
##       airline_code=OO}         => {airline_name=Northwest} 0.01047630          1 9.382067  1356
## [17] {satisfaction=(1,2],                                                                      
##       gender=Female,                                                                           
##       type_of_travel=Personal,                                                                 
##       airline_name=Northwest}  => {airline_code=OO}        0.01002047          1 9.382067  1297
## [18] {satisfaction=(1,2],                                                                      
##       gender=Female,                                                                           
##       type_of_travel=Personal,                                                                 
##       airline_code=OO}         => {airline_name=Northwest} 0.01002047          1 9.382067  1297

• 52 rules (sorted decreasing by count, confidence = .8, lift > 8) were generated for female travelers (rating = 2)
• the top 20 rules (sorted by decreasing count) were for airlines: FlyFast and Sigma with airline code: EV and DL

Let’s dig deeper into why females give higher ratings (4)

female_ruleset <- apriori(dfX, parameter=list(support=.1, confidence=.5, minlen=2, maxlen=6, maxtime=20))

## Apriori
## 
## Parameter specification:
##  confidence minval smax arem  aval originalSupport maxtime support minlen
##         0.5    0.1    1 none FALSE            TRUE      20     0.1      2
##  maxlen target   ext
##       6  rules FALSE
## 
## Algorithmic control:
##  filter tree heap memopt load sort verbose
##     0.1 TRUE TRUE  FALSE TRUE    2    TRUE
## 
## Absolute minimum support count: 12943 
## 
## set item appearances ...[0 item(s)] done [0.00s].
## set transactions ...[896 item(s), 129435 transaction(s)] done [0.42s].
## sorting and recoding items ... [73 item(s)] done [0.04s].
## creating transaction tree ... done [0.09s].
## checking subsets of size 1 2 3 4 5 6

## Warning in apriori(dfX, parameter = list(support = 0.1, confidence =
## 0.5, : Mining stopped (maxlen reached). Only patterns up to a length of 6
## returned!

##  done [1.10s].
## writing ... [13955 rule(s)] done [0.00s].
## creating S4 object  ... done [0.03s].

female_filter <- subset(female_ruleset, subset=lhs %ain% c("gender=Female","satisfaction=(3,4]") & lift>1.1)
plot(female_filter)

female_filter <- sort(female_filter, decreasing=T, by="count")
summary(female_filter)

## set of 53 rules
## 
## rule length distribution (lhs + rhs):sizes
##  3  4  5  6 
##  2 13 24 14 
## 
##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##   3.000   4.000   5.000   4.943   6.000   6.000 
## 
## summary of quality measures:
##     support         confidence          lift           count      
##  Min.   :0.1006   Min.   :0.7295   Min.   :1.110   Min.   :13017  
##  1st Qu.:0.1136   1st Qu.:0.7930   1st Qu.:1.256   1st Qu.:14698  
##  Median :0.1289   Median :0.8210   Median :1.303   Median :16689  
##  Mean   :0.1287   Mean   :0.8423   Mean   :1.273   Mean   :16656  
##  3rd Qu.:0.1365   3rd Qu.:0.8811   3rd Qu.:1.336   3rd Qu.:17662  
##  Max.   :0.1861   Max.   :0.9792   Max.   :1.399   Max.   :24090  
## 
## mining info:
##  data ntransactions support confidence
##   dfX        129435     0.1        0.5

inspect(female_filter[1:20,])

##      lhs                                    rhs                                   support confidence     lift count
## [1]  {satisfaction=(3,4],                                                                                          
##       gender=Female}                     => {type_of_travel=Business}           0.1861166  0.8215394 1.339497 24090
## [2]  {satisfaction=(3,4],                                                                                          
##       gender=Female,                                                                                               
##       flight_cancelled=No}               => {type_of_travel=Business}           0.1836057  0.8242006 1.343836 23765
## [3]  {satisfaction=(3,4],                                                                                          
##       gender=Female}                     => {arrival_delay_greater_5_mins=no}   0.1707421  0.7536746 1.146671 22100
## [4]  {satisfaction=(3,4],                                                                                          
##       gender=Female,                                                                                               
##       arrival_delay_greater_5_mins=no}   => {departure_delay_greater_5_mins=no} 0.1671186  0.9787783 1.256914 21631
## [5]  {satisfaction=(3,4],                                                                                          
##       gender=Female,                                                                                               
##       departure_delay_greater_5_mins=no} => {arrival_delay_greater_5_mins=no}   0.1671186  0.8817822 1.341579 21631
## [6]  {satisfaction=(3,4],                                                                                          
##       gender=Female,                                                                                               
##       flight_cancelled=No}               => {arrival_delay_greater_5_mins=no}   0.1669641  0.7494971 1.140315 21611
## [7]  {satisfaction=(3,4],                                                                                          
##       gender=Female,                                                                                               
##       flight_cancelled=No,                                                                                         
##       arrival_delay_greater_5_mins=no}   => {departure_delay_greater_5_mins=no} 0.1633639  0.9784369 1.256476 21145
## [8]  {satisfaction=(3,4],                                                                                          
##       gender=Female,                                                                                               
##       departure_delay_greater_5_mins=no,                                                                           
##       flight_cancelled=No}               => {arrival_delay_greater_5_mins=no}   0.1633639  0.8793928 1.337944 21145
## [9]  {satisfaction=(3,4],                                                                                          
##       gender=Female,                                                                                               
##       departure_delay_greater_5_mins=no} => {type_of_travel=Business}           0.1534361  0.8095879 1.320010 19860
## [10] {satisfaction=(3,4],                                                                                          
##       gender=Female,                                                                                               
##       departure_delay_greater_5_mins=no,                                                                           
##       flight_cancelled=No}               => {type_of_travel=Business}           0.1509484  0.8125598 1.324856 19538
## [11] {satisfaction=(3,4],                                                                                          
##       gender=Female,                                                                                               
##       class=Economy}                     => {type_of_travel=Business}           0.1421640  0.8159365 1.330361 18401
## [12] {satisfaction=(3,4],                                                                                          
##       gender=Female,                                                                                               
##       class=Economy,                                                                                               
##       flight_cancelled=No}               => {type_of_travel=Business}           0.1400626  0.8189827 1.335328 18129
## [13] {satisfaction=(3,4],                                                                                          
##       gender=Female,                                                                                               
##       type_of_travel=Business}           => {arrival_delay_greater_5_mins=no}   0.1364546  0.7331673 1.115470 17662
## [14] {satisfaction=(3,4],                                                                                          
##       gender=Female,                                                                                               
##       arrival_delay_greater_5_mins=no}   => {type_of_travel=Business}           0.1364546  0.7991855 1.303049 17662
## [15] {satisfaction=(3,4],                                                                                          
##       gender=Female,                                                                                               
##       type_of_travel=Business,                                                                                     
##       flight_cancelled=No}               => {arrival_delay_greater_5_mins=no}   0.1339437  0.7295182 1.109918 17337
## [16] {satisfaction=(3,4],                                                                                          
##       gender=Female,                                                                                               
##       flight_cancelled=No,                                                                                         
##       arrival_delay_greater_5_mins=no}   => {type_of_travel=Business}           0.1339437  0.8022303 1.308014 17337
## [17] {satisfaction=(3,4],                                                                                          
##       gender=Female,                                                                                               
##       type_of_travel=Business,                                                                                     
##       arrival_delay_greater_5_mins=no}   => {departure_delay_greater_5_mins=no} 0.1335806  0.9789378 1.257119 17290
## [18] {satisfaction=(3,4],                                                                                          
##       gender=Female,                                                                                               
##       type_of_travel=Business,                                                                                     
##       departure_delay_greater_5_mins=no} => {arrival_delay_greater_5_mins=no}   0.1335806  0.8705942 1.324557 17290
## [19] {satisfaction=(3,4],                                                                                          
##       gender=Female,                                                                                               
##       departure_delay_greater_5_mins=no,                                                                           
##       arrival_delay_greater_5_mins=no}   => {type_of_travel=Business}           0.1335806  0.7993158 1.303262 17290
## [20] {satisfaction=(3,4],                                                                                          
##       gender=Female,                                                                                               
##       class=Economy}                     => {arrival_delay_greater_5_mins=no}   0.1318345  0.7566513 1.151200 17064

• 53 rules were generated; the top 20 rules (confidence > .8, lift > 1.2) were for female business travelers with delay < 5 mins.
• No personal travel was overserved for this subset of the rules (1-30)
• Business travel was the number 1 rule with 24090 observations

Let’s dig deeper into why females give higher ratings (5)

female_ruleset <- apriori(dfX, parameter=list(support=.01, confidence=.5, minlen=3, maxlen=6, maxtime=10))

## Apriori
## 
## Parameter specification:
##  confidence minval smax arem  aval originalSupport maxtime support minlen
##         0.5    0.1    1 none FALSE            TRUE      10    0.01      3
##  maxlen target   ext
##       6  rules FALSE
## 
## Algorithmic control:
##  filter tree heap memopt load sort verbose
##     0.1 TRUE TRUE  FALSE TRUE    2    TRUE
## 
## Absolute minimum support count: 1294 
## 
## set item appearances ...[0 item(s)] done [0.00s].
## set transactions ...[896 item(s), 129435 transaction(s)] done [0.44s].
## sorting and recoding items ... [282 item(s)] done [0.05s].
## creating transaction tree ... done [0.11s].
## checking subsets of size 1 2 3 4 5

## Warning in apriori(dfX, parameter = list(support = 0.01, confidence =
## 0.5, : Mining stopped (time limit reached). Only patterns up to a length of
## 5 returned!

##  done [12.94s].
## writing ... [1335284 rule(s)] done [0.12s].
## creating S4 object  ... done [0.22s].

female_filter <- subset(female_ruleset, subset=lhs %ain% c("gender=Female","satisfaction=(4,5]") & lift>1.5)
plot(female_filter)

## To reduce overplotting, jitter is added! Use jitter = 0 to prevent jitter.

female_filter <- sort(female_filter, decreasing=T, by="count")
inspect(female_filter[1:20,])

##      lhs                                    rhs                          support confidence     lift count
## [1]  {satisfaction=(4,5],                                                                                 
##       gender=Female}                     => {type_of_travel=Business} 0.03025457  0.9317154 1.519136  3916
## [2]  {satisfaction=(4,5],                                                                                 
##       gender=Female,                                                                                      
##       flight_cancelled=No}               => {type_of_travel=Business} 0.03013868  0.9316933 1.519100  3901
## [3]  {satisfaction=(4,5],                                                                                 
##       gender=Female,                                                                                      
##       class=Economy}                     => {type_of_travel=Business} 0.02320856  0.9291680 1.514982  3004
## [4]  {satisfaction=(4,5],                                                                                 
##       gender=Female,                                                                                      
##       class=Economy,                                                                                      
##       flight_cancelled=No}               => {type_of_travel=Business} 0.02311585  0.9291925 1.515022  2992
## [5]  {satisfaction=(4,5],                                                                                 
##       gender=Female,                                                                                      
##       price_sensitivity=(0,1]}           => {type_of_travel=Business} 0.02309267  0.9308627 1.517745  2989
## [6]  {satisfaction=(4,5],                                                                                 
##       gender=Female,                                                                                      
##       departure_delay_greater_5_mins=no} => {type_of_travel=Business} 0.02304632  0.9269733 1.511404  2983
## [7]  {satisfaction=(4,5],                                                                                 
##       gender=Female,                                                                                      
##       price_sensitivity=(0,1],                                                                            
##       flight_cancelled=No}               => {type_of_travel=Business} 0.02299996  0.9308943 1.517797  2977
## [8]  {satisfaction=(4,5],                                                                                 
##       gender=Female,                                                                                      
##       departure_delay_greater_5_mins=no,                                                                  
##       flight_cancelled=No}               => {type_of_travel=Business} 0.02293043  0.9269207 1.511318  2968
## [9]  {satisfaction=(4,5],                                                                                 
##       gender=Female,                                                                                      
##       arrival_delay_greater_5_mins=no}   => {type_of_travel=Business} 0.01952331  0.9259802 1.509785  2527
## [10] {satisfaction=(4,5],                                                                                 
##       gender=Female,                                                                                      
##       flight_cancelled=No,                                                                                
##       arrival_delay_greater_5_mins=no}   => {type_of_travel=Business} 0.01940742  0.9259123 1.509674  2512
## [11] {satisfaction=(4,5],                                                                                 
##       gender=Female,                                                                                      
##       departure_delay_greater_5_mins=no,                                                                  
##       arrival_delay_greater_5_mins=no}   => {type_of_travel=Business} 0.01909839  0.9254961 1.508995  2472
## [12] {satisfaction=(4,5],                                                                                 
##       gender=Female,                                                                                      
##       class=Economy,                                                                                      
##       departure_delay_greater_5_mins=no} => {type_of_travel=Business} 0.01783134  0.9257922 1.509478  2308
## [13] {satisfaction=(4,5],                                                                                 
##       gender=Female,                                                                                      
##       price_sensitivity=(0,1],                                                                            
##       departure_delay_greater_5_mins=no} => {type_of_travel=Business} 0.01764592  0.9265720 1.510749  2284
## [14] {satisfaction=(4,5],                                                                                 
##       gender=Female,                                                                                      
##       price_sensitivity=(0,1],                                                                            
##       class=Economy}                     => {type_of_travel=Business} 0.01750686  0.9290693 1.514821  2266
## [15] {satisfaction=(4,5],                                                                                 
##       gender=Female,                                                                                      
##       class=Economy,                                                                                      
##       arrival_delay_greater_5_mins=no}   => {type_of_travel=Business} 0.01511183  0.9256981 1.509325  1956
## [16] {satisfaction=(4,5],                                                                                 
##       gender=Female,                                                                                      
##       num_loyalty_cards=(-1,0]}          => {type_of_travel=Business} 0.01508865  0.9335564 1.522137  1953
## [17] {satisfaction=(4,5],                                                                                 
##       gender=Female,                                                                                      
##       num_loyalty_cards=(-1,0],                                                                           
##       flight_cancelled=No}               => {type_of_travel=Business} 0.01503457  0.9337812 1.522504  1946
## [18] {satisfaction=(4,5],                                                                                 
##       gender=Female,                                                                                      
##       price_sensitivity=(0,1],                                                                            
##       arrival_delay_greater_5_mins=no}   => {type_of_travel=Business} 0.01502685  0.9253092 1.508691  1945
## [19] {satisfaction=(4,5],                                                                                 
##       gender=Female,                                                                                      
##       airport_shopping=(-1,0]}           => {type_of_travel=Business} 0.01453239  0.9238703 1.506344  1881
## [20] {satisfaction=(4,5],                                                                                 
##       gender=Female,                                                                                      
##       airport_shopping=(-1,0],                                                                            
##       flight_cancelled=No}               => {type_of_travel=Business} 0.01448604  0.9236453 1.505978  1875

• Of the 54 rules returned (confidence > .9, lift > 1.5), all of them include female travelers with the following characteristics: type_of_travel=Business
• Surprisingly, class=Economy shows up as the highest association with/without price_sensitivity=1 and satisfaction=5
• Unsurprisingly, arrival_delay_greater_5_mins = no is a common pairing in the rules
• This is the first ruleset where age is a common pairing. Female, age groups of 34-46 and 46-58 are associated with business travel and rating=5
• Silver status was observed in two rules. No Blue status were observed in any of the rules

get ruleset to only show associations for females age(46-58) with ratings of 4

ruleset <- apriori(dfX, parameter=list(support=.01, confidence=.5, minlen=2, maxlen=6, maxtime=20))

## Apriori
## 
## Parameter specification:
##  confidence minval smax arem  aval originalSupport maxtime support minlen
##         0.5    0.1    1 none FALSE            TRUE      20    0.01      2
##  maxlen target   ext
##       6  rules FALSE
## 
## Algorithmic control:
##  filter tree heap memopt load sort verbose
##     0.1 TRUE TRUE  FALSE TRUE    2    TRUE
## 
## Absolute minimum support count: 1294 
## 
## set item appearances ...[0 item(s)] done [0.00s].
## set transactions ...[896 item(s), 129435 transaction(s)] done [0.41s].
## sorting and recoding items ... [282 item(s)] done [0.05s].
## creating transaction tree ... done [0.10s].
## checking subsets of size 1 2 3 4 5 6

## Warning in apriori(dfX, parameter = list(support = 0.01, confidence =
## 0.5, : Mining stopped (maxlen reached). Only patterns up to a length of 6
## returned!

##  done [30.43s].
## writing ... [2741159 rule(s)] done [0.31s].
## creating S4 object  ... done [0.70s].

female_rules <- subset(female_ruleset, subset=lhs %ain% c("gender=Female","satisfaction=(3,4]","age=(46,58]") & lift>1.5)
plot(female_rules)

## To reduce overplotting, jitter is added! Use jitter = 0 to prevent jitter.

female_rules <- sort(female_rules, decreasing = T, by="count")
inspect(female_rules)

##     lhs                                      rhs                                      support confidence     lift count
## [1] {satisfaction=(3,4],                                                                                               
##      age=(46,58],                                                                                                      
##      gender=Female,                                                                                                    
##      percent_flight_other_airlines=(0,4]} => {num_loyalty_cards=(-1,0]}            0.01551358  0.8243021 1.555753  2008
## [2] {satisfaction=(3,4],                                                                                               
##      age=(46,58],                                                                                                      
##      gender=Female,                                                                                                    
##      flight_distance=(32,363]}            => {flight_time_in_minutes=(7,61]}       0.01333488  0.9239829 3.487874  1726
## [3] {satisfaction=(3,4],                                                                                               
##      age=(46,58],                                                                                                      
##      gender=Female,                                                                                                    
##      flight_time_in_minutes=(7,61]}       => {flight_distance=(32,363]}            0.01333488  0.8621379 3.428816  1726
## [4] {satisfaction=(3,4],                                                                                               
##      age=(46,58],                                                                                                      
##      gender=Female,                                                                                                    
##      flight_time_in_minutes=(142,670]}    => {flight_distance=(1.02e+03,4.98e+03]} 0.01219917  0.8733407 3.509387  1579
## [5] {satisfaction=(3,4],                                                                                               
##      age=(46,58],                                                                                                      
##      gender=Female,                                                                                                    
##      flight_distance=(1.02e+03,4.98e+03]} => {flight_time_in_minutes=(142,670]}    0.01219917  0.8675824 3.565956  1579
## [6] {satisfaction=(3,4],                                                                                               
##      age=(46,58],                                                                                                      
##      gender=Female,                                                                                                    
##      airline_code=WN}                     => {airline_name=Cheapseats}             0.01093985  1.0000000 4.983828  1416
## [7] {satisfaction=(3,4],                                                                                               
##      age=(46,58],                                                                                                      
##      gender=Female,                                                                                                    
##      airline_name=Cheapseats}             => {airline_code=WN}                     0.01093985  1.0000000 4.983828  1416
## [8] {satisfaction=(3,4],                                                                                               
##      age=(46,58],                                                                                                      
##      gender=Female,                                                                                                    
##      flight_time_in_minutes=(61,95]}      => {flight_distance=(363,630]}           0.01039132  0.7756632 3.103684  1345
## [9] {satisfaction=(3,4],                                                                                               
##      age=(46,58],                                                                                                      
##      gender=Female,                                                                                                    
##      flight_distance=(363,630]}           => {flight_time_in_minutes=(61,95]}      0.01039132  0.7509771 3.063047  1345

• 4 rules show with high confidence (1.0) and lift (4.98) that females, age 46-59 flying with Cheapseats on airline code WN give a satisfaction rating of 4
• Surprisingly, the top 3 of the 32 rules returned were for female business travelers, age 46-58, and 0 loyalty cards
• These travelors also seem to either take shorter flights that are about an hour in length, or longer flights that are up to 12 hours

get ruleset for females age(34_46) with rating 4

ruleset <- apriori(dfX, parameter=list(support=.05, confidence=.5, minlen=2, maxlen=6, maxtime=20))

## Apriori
## 
## Parameter specification:
##  confidence minval smax arem  aval originalSupport maxtime support minlen
##         0.5    0.1    1 none FALSE            TRUE      20    0.05      2
##  maxlen target   ext
##       6  rules FALSE
## 
## Algorithmic control:
##  filter tree heap memopt load sort verbose
##     0.1 TRUE TRUE  FALSE TRUE    2    TRUE
## 
## Absolute minimum support count: 6471 
## 
## set item appearances ...[0 item(s)] done [0.00s].
## set transactions ...[896 item(s), 129435 transaction(s)] done [0.41s].
## sorting and recoding items ... [98 item(s)] done [0.05s].
## creating transaction tree ... done [0.09s].
## checking subsets of size 1 2 3 4 5 6

## Warning in apriori(dfX, parameter = list(support = 0.05, confidence =
## 0.5, : Mining stopped (maxlen reached). Only patterns up to a length of 6
## returned!

##  done [3.75s].
## writing ... [87030 rule(s)] done [0.01s].
## creating S4 object  ... done [0.05s].

female_filter <- subset(ruleset, subset=lhs %ain% c("gender=Female","satisfaction=(3,4]","age=(34,46]") & lift>1)
plot(female_filter)

female_filter <- sort(female_filter, decreasing = T, by="lift")
summary(female_filter)

## set of 19 rules
## 
## rule length distribution (lhs + rhs):sizes
##  4  5  6 
##  5 11  3 
## 
##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##   4.000   4.500   5.000   4.895   5.000   6.000 
## 
## summary of quality measures:
##     support          confidence          lift           count     
##  Min.   :0.05023   Min.   :0.7167   Min.   :1.000   Min.   :6501  
##  1st Qu.:0.05086   1st Qu.:0.8211   1st Qu.:1.006   1st Qu.:6582  
##  Median :0.05140   Median :0.8712   Median :1.057   Median :6653  
##  Mean   :0.05481   Mean   :0.8858   Mean   :1.127   Mean   :7095  
##  3rd Qu.:0.05826   3rd Qu.:0.9831   3rd Qu.:1.255   3rd Qu.:7540  
##  Max.   :0.07048   Max.   :0.9904   Max.   :1.388   Max.   :9123  
## 
## mining info:
##  data ntransactions support confidence
##   dfX        129435    0.05        0.5

inspect(female_filter)

##      lhs                                    rhs                                    support confidence     lift count
## [1]  {satisfaction=(3,4],                                                                                           
##       age=(34,46],                                                                                                  
##       gender=Female,                                                                                                
##       flight_cancelled=No}               => {type_of_travel=Business}           0.05999150  0.8511455 1.387769  7765
## [2]  {satisfaction=(3,4],                                                                                           
##       age=(34,46],                                                                                                  
##       gender=Female}                     => {type_of_travel=Business}           0.06057094  0.8481177 1.382832  7840
## [3]  {satisfaction=(3,4],                                                                                           
##       age=(34,46],                                                                                                  
##       gender=Female,                                                                                                
##       departure_delay_greater_5_mins=no} => {arrival_delay_greater_5_mins=no}   0.05116081  0.8712012 1.325480  6622
## [4]  {satisfaction=(3,4],                                                                                           
##       age=(34,46],                                                                                                  
##       gender=Female,                                                                                                
##       departure_delay_greater_5_mins=no,                                                                            
##       flight_cancelled=No}               => {arrival_delay_greater_5_mins=no}   0.05022598  0.8691176 1.322310  6501
## [5]  {satisfaction=(3,4],                                                                                           
##       age=(34,46],                                                                                                  
##       gender=Female,                                                                                                
##       arrival_delay_greater_5_mins=no}   => {departure_delay_greater_5_mins=no} 0.05116081  0.9775613 1.255351  6622
## [6]  {satisfaction=(3,4],                                                                                           
##       age=(34,46],                                                                                                  
##       gender=Female,                                                                                                
##       flight_cancelled=No,                                                                                          
##       arrival_delay_greater_5_mins=no}   => {departure_delay_greater_5_mins=no} 0.05022598  0.9771532 1.254827  6501
## [7]  {satisfaction=(3,4],                                                                                           
##       age=(34,46],                                                                                                  
##       gender=Female}                     => {arrival_delay_greater_5_mins=no}   0.05233515  0.7327997 1.114911  6774
## [8]  {satisfaction=(3,4],                                                                                           
##       age=(34,46],                                                                                                  
##       gender=Female,                                                                                                
##       flight_cancelled=No}               => {arrival_delay_greater_5_mins=no}   0.05140032  0.7292557 1.109519  6653
## [9]  {satisfaction=(3,4],                                                                                           
##       age=(34,46],                                                                                                  
##       gender=Female,                                                                                                
##       flight_cancelled=No}               => {price_sensitivity=(0,1]}           0.05055047  0.7171983 1.058055  6543
## [10] {satisfaction=(3,4],                                                                                           
##       age=(34,46],                                                                                                  
##       gender=Female}                     => {price_sensitivity=(0,1]}           0.05118399  0.7166811 1.057292  6625
## [11] {satisfaction=(3,4],                                                                                           
##       age=(34,46],                                                                                                  
##       gender=Female}                     => {departure_delay_greater_5_mins=no} 0.05872446  0.8222631 1.055923  7601
## [12] {satisfaction=(3,4],                                                                                           
##       age=(34,46],                                                                                                  
##       gender=Female,                                                                                                
##       flight_cancelled=No}               => {departure_delay_greater_5_mins=no} 0.05778962  0.8199057 1.052896  7480
## [13] {satisfaction=(3,4],                                                                                           
##       age=(34,46],                                                                                                  
##       gender=Female,                                                                                                
##       type_of_travel=Business}           => {flight_cancelled=No}               0.05999150  0.9904337 1.009066  7765
## [14] {satisfaction=(3,4],                                                                                           
##       age=(34,46],                                                                                                  
##       gender=Female,                                                                                                
##       price_sensitivity=(0,1]}           => {flight_cancelled=No}               0.05055047  0.9876226 1.006202  6543
## [15] {satisfaction=(3,4],                                                                                           
##       age=(34,46],                                                                                                  
##       gender=Female}                     => {flight_cancelled=No}               0.07048325  0.9869104 1.005476  9123
## [16] {satisfaction=(3,4],                                                                                           
##       age=(34,46],                                                                                                  
##       gender=Female,                                                                                                
##       class=Economy}                     => {flight_cancelled=No}               0.05569591  0.9853745 1.003912  7209
## [17] {satisfaction=(3,4],                                                                                           
##       age=(34,46],                                                                                                  
##       gender=Female,                                                                                                
##       departure_delay_greater_5_mins=no} => {flight_cancelled=No}               0.05778962  0.9840810 1.002594  7480
## [18] {satisfaction=(3,4],                                                                                           
##       age=(34,46],                                                                                                  
##       gender=Female,                                                                                                
##       arrival_delay_greater_5_mins=no}   => {flight_cancelled=No}               0.05140032  0.9821376 1.000614  6653
## [19] {satisfaction=(3,4],                                                                                           
##       age=(34,46],                                                                                                  
##       gender=Female,                                                                                                
##       departure_delay_greater_5_mins=no,                                                                            
##       arrival_delay_greater_5_mins=no}   => {flight_cancelled=No}               0.05022598  0.9817276 1.000196  6501

• No surprises here. This analysis shows that female business travelers, age 34-46, with no delays or cancellations, and price_sensitity=1 give ratings=4

get ruleset to only show associations for females age(46-58) with ratings of 5

female_ruleset <- apriori(dfX, parameter=list(support=.01, confidence=.5, minlen=2, maxlen=6, maxtime=20))

## Apriori
## 
## Parameter specification:
##  confidence minval smax arem  aval originalSupport maxtime support minlen
##         0.5    0.1    1 none FALSE            TRUE      20    0.01      2
##  maxlen target   ext
##       6  rules FALSE
## 
## Algorithmic control:
##  filter tree heap memopt load sort verbose
##     0.1 TRUE TRUE  FALSE TRUE    2    TRUE
## 
## Absolute minimum support count: 1294 
## 
## set item appearances ...[0 item(s)] done [0.00s].
## set transactions ...[896 item(s), 129435 transaction(s)] done [0.46s].
## sorting and recoding items ... [282 item(s)] done [0.06s].
## creating transaction tree ... done [0.11s].
## checking subsets of size 1 2 3 4 5 6

## Warning in apriori(dfX, parameter = list(support = 0.01, confidence =
## 0.5, : Mining stopped (maxlen reached). Only patterns up to a length of 6
## returned!

##  done [30.85s].
## writing ... [2741159 rule(s)] done [0.26s].
## creating S4 object  ... done [0.48s].

female_filter <- subset(female_ruleset, subset=lhs %ain% c("gender=Female","satisfaction=(4,5]","age=(46,58]") & lift>1)
female_filter <- sort(female_filter, decreasing = T, by="lift")
summary(female_filter)

## set of 4 rules
## 
## rule length distribution (lhs + rhs):sizes
## 4 5 
## 2 2 
## 
##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##     4.0     4.0     4.5     4.5     5.0     5.0 
## 
## summary of quality measures:
##     support          confidence          lift           count     
##  Min.   :0.01018   Min.   :0.9341   Min.   :1.014   Min.   :1318  
##  1st Qu.:0.01018   1st Qu.:0.9343   1st Qu.:1.014   1st Qu.:1318  
##  Median :0.01021   Median :0.9649   Median :1.269   Median :1321  
##  Mean   :0.01037   Mean   :0.9649   Mean   :1.269   Mean   :1343  
##  3rd Qu.:0.01040   3rd Qu.:0.9955   3rd Qu.:1.523   3rd Qu.:1346  
##  Max.   :0.01090   Max.   :0.9958   Max.   :1.523   Max.   :1411  
## 
## mining info:
##  data ntransactions support confidence
##   dfX        129435    0.01        0.5

inspect(female_filter)

##     lhs                          rhs                          support confidence     lift count
## [1] {satisfaction=(4,5],                                                                       
##      age=(46,58],                                                                              
##      gender=Female}           => {type_of_travel=Business} 0.01022907  0.9343684 1.523461  1324
## [2] {satisfaction=(4,5],                                                                       
##      age=(46,58],                                                                              
##      gender=Female,                                                                            
##      flight_cancelled=No}     => {type_of_travel=Business} 0.01018272  0.9340893 1.523006  1318
## [3] {satisfaction=(4,5],                                                                       
##      age=(46,58],                                                                              
##      gender=Female}           => {flight_cancelled=No}     0.01090122  0.9957657 1.014498  1411
## [4] {satisfaction=(4,5],                                                                       
##      age=(46,58],                                                                              
##      gender=Female,                                                                            
##      type_of_travel=Business} => {flight_cancelled=No}     0.01018272  0.9954683 1.014195  1318

• This analysis again shows that female business travelers, age 46-59 on flights that are on-time give high ratings

get ruleset for females age(34-46) with rating 5

female_ruleset <- apriori(dfX, parameter=list(support=.01, confidence=.5, minlen=2, maxlen=6, maxtime=20))

## Apriori
## 
## Parameter specification:
##  confidence minval smax arem  aval originalSupport maxtime support minlen
##         0.5    0.1    1 none FALSE            TRUE      20    0.01      2
##  maxlen target   ext
##       6  rules FALSE
## 
## Algorithmic control:
##  filter tree heap memopt load sort verbose
##     0.1 TRUE TRUE  FALSE TRUE    2    TRUE
## 
## Absolute minimum support count: 1294 
## 
## set item appearances ...[0 item(s)] done [0.00s].
## set transactions ...[896 item(s), 129435 transaction(s)] done [0.42s].
## sorting and recoding items ... [282 item(s)] done [0.05s].
## creating transaction tree ... done [0.11s].
## checking subsets of size 1 2 3 4 5 6

## Warning in apriori(dfX, parameter = list(support = 0.01, confidence =
## 0.5, : Mining stopped (maxlen reached). Only patterns up to a length of 6
## returned!

##  done [30.59s].
## writing ... [2741159 rule(s)] done [0.32s].
## creating S4 object  ... done [0.51s].

female_filter <- subset(female_ruleset, subset=lhs %ain% c("gender=Female","satisfaction=(4,5]","age=(34,46]") & lift>1)
female_filter <- sort(female_filter, decreasing = T, by="lift")
summary(female_filter)

## set of 9 rules
## 
## rule length distribution (lhs + rhs):sizes
## 4 5 
## 3 6 
## 
##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##   4.000   4.000   5.000   4.667   5.000   5.000 
## 
## summary of quality measures:
##     support          confidence          lift           count     
##  Min.   :0.01045   Min.   :0.7822   Min.   :1.014   Min.   :1352  
##  1st Qu.:0.01062   1st Qu.:0.9257   1st Qu.:1.015   1st Qu.:1375  
##  Median :0.01065   Median :0.9956   Median :1.016   Median :1379  
##  Mean   :0.01158   Mean   :0.9331   Mean   :1.156   Mean   :1499  
##  3rd Qu.:0.01257   3rd Qu.:0.9966   3rd Qu.:1.155   3rd Qu.:1627  
##  Max.   :0.01357   Max.   :0.9971   Max.   :1.510   Max.   :1757  
## 
## mining info:
##  data ntransactions support confidence
##   dfX        129435    0.01        0.5

inspect(female_filter)

##     lhs                                    rhs                          support confidence     lift count
## [1] {satisfaction=(4,5],                                                                                 
##      age=(34,46],                                                                                        
##      gender=Female,                                                                                      
##      flight_cancelled=No}               => {type_of_travel=Business} 0.01257002  0.9260102 1.509834  1627
## [2] {satisfaction=(4,5],                                                                                 
##      age=(34,46],                                                                                        
##      gender=Female}                     => {type_of_travel=Business} 0.01260865  0.9256948 1.509319  1632
## [3] {satisfaction=(4,5],                                                                                 
##      age=(34,46],                                                                                        
##      gender=Female,                                                                                      
##      flight_cancelled=No}               => {price_sensitivity=(0,1]} 0.01062309  0.7825839 1.154516  1375
## [4] {satisfaction=(4,5],                                                                                 
##      age=(34,46],                                                                                        
##      gender=Female}                     => {price_sensitivity=(0,1]} 0.01065400  0.7821894 1.153934  1379
## [5] {satisfaction=(4,5],                                                                                 
##      age=(34,46],                                                                                        
##      gender=Female,                                                                                      
##      price_sensitivity=(0,1]}           => {flight_cancelled=No}     0.01062309  0.9970993 1.015857  1375
## [6] {satisfaction=(4,5],                                                                                 
##      age=(34,46],                                                                                        
##      gender=Female,                                                                                      
##      type_of_travel=Business}           => {flight_cancelled=No}     0.01257002  0.9969363 1.015691  1627
## [7] {satisfaction=(4,5],                                                                                 
##      age=(34,46],                                                                                        
##      gender=Female}                     => {flight_cancelled=No}     0.01357438  0.9965967 1.015345  1757
## [8] {satisfaction=(4,5],                                                                                 
##      age=(34,46],                                                                                        
##      gender=Female,                                                                                      
##      class=Economy}                     => {flight_cancelled=No}     0.01054583  0.9956236 1.014354  1365
## [9] {satisfaction=(4,5],                                                                                 
##      age=(34,46],                                                                                        
##      gender=Female,                                                                                      
##      departure_delay_greater_5_mins=no} => {flight_cancelled=No}     0.01044540  0.9955817 1.014311  1352

• Very similar to preceding analysis: female business travelers, economy class, price_sensitivity=1 give rating=5