Facebook “Sample” Birthday Investigation

Note: I am using the sample data because I am worried about invasion of privacy using real Facebook data.

bd <- read.csv("../birthdaysExample.csv", header = TRUE)
str(bd)
## 'data.frame':    1033 obs. of  1 variable:
##  $ dates: Factor w/ 348 levels "1/1/14","1/10/14",..: 78 258 322 219 131 241 33 46 287 331 ...
plot(bd, xlab = "Birthdays", ylab = "Count", main = "Test Birthday Plot")

We start out just seeing what kind of data is there - a simple plot shows the dates are not in date order but are in string order… Note that the given birthdays are all in the same year, so no age distributions can be done. The maximum number of birthdays on a given day is 8 and there are 3 days that are at the maximum.

Need to convert factors to actual dates, and will be convenient to add month and day columns:

bd$ddates <- strptime(bd$dates, format = "%m/%d/%y")
bd$dates <- as.Date(bd$dates, format = "%m/%d/%y")
bd$months <- as.numeric(strftime(bd$dates, format = "%m"))
bd$days <- as.numeric(strftime(bd$dates, "%d"))
head(bd)
##        dates     ddates months days
## 1 2014-11-25 2014-11-25     11   25
## 2 2014-06-08 2014-06-08      6    8
## 3 2014-09-12 2014-09-12      9   12
## 4 2014-05-26 2014-05-26      5   26
## 5 2014-02-20 2014-02-20      2   20
## 6 2014-06-19 2014-06-19      6   19

Let’s look at a summary of the birthdays…

summary(bd)
##      dates                ddates                        months      
##  Min.   :2014-01-01   Min.   :2014-01-01 00:00:00   Min.   : 1.000  
##  1st Qu.:2014-03-28   1st Qu.:2014-03-28 00:00:00   1st Qu.: 3.000  
##  Median :2014-07-02   Median :2014-07-02 00:00:00   Median : 7.000  
##  Mean   :2014-06-30   Mean   :2014-06-30 10:19:27   Mean   : 6.474  
##  3rd Qu.:2014-09-28   3rd Qu.:2014-09-28 00:00:00   3rd Qu.: 9.000  
##  Max.   :2014-12-31   Max.   :2014-12-31 00:00:00   Max.   :12.000  
##       days     
##  Min.   : 1.0  
##  1st Qu.: 8.0  
##  Median :16.0  
##  Mean   :15.7  
##  3rd Qu.:23.0  
##  Max.   :31.0

I don’t want to take up space here with a table of birthdays:
table(bd$dates)

How many people share your birthday? Do you know them?

2 people share my birthday (do not know who they are since this is sample data)

which(bd$dates == mybd)
## [1]  18 893

Which month contains the most number of birthdays?

March (98)

monthTable <- table(bd$months)
names(monthTable) <- month.name
which(monthTable == max(monthTable))
## March 
##     3

How many birthdays are in each month?

table(bd$months)
## 
##  1  2  3  4  5  6  7  8  9 10 11 12 
## 89 79 98 81 72 93 86 91 96 89 87 72
# does not have month names
monthTable
##   January  February     March     April       May      June      July 
##        89        79        98        81        72        93        86 
##    August September   October  November  December 
##        91        96        89        87        72
# has month names - simplest and more complex
plot(monthTable)

ggplot(data = bd, aes(x=as.factor(months))) + geom_histogram(colour = "darkgreen", fill = "white", binwidth = 1 ) +
  scale_x_discrete( labels = month.name)+ theme(axis.text.x = element_text(angle = 90, hjust = 1)) + labs(title = "Birthday Count by Month") + xlab("Month")

Which day of the year has the most number of birthdays?

We saw from the initial plot that the maximum number of birthdays per day is 8

dateTable <- table(bd$dates)
which(dateTable ==max(dateTable) )
## 2014-02-06 2014-05-22 2014-07-16 
##         37        135        188

Do you have at least 365 friends that have birthdays on everyday of the year?

no
#http://stackoverflow.com/questions/1330989/rotating-and-spacing-axis-labels-in-ggplot2

#Convert day of year to date format
#https  ://stat.ethz.ch/pipermail/r-help/2012-March/308013.html
bt <- table (bd$dates)
length(names(bt))
## [1] 348
missing = 365 - length(names(bt))
missing
## [1] 17
all_days <- lapply(as.array(seq(0,364)), as.Date, origin = "2014-01-01" )
datefun2 <- function(x) {if (is.na(bt[as.character(x)])) x}
nulldates <- lapply(all_days,datefun2)
onlynulldates <- nulldates[!sapply(nulldates, is.null)]
# Help from https://stat.ethz.ch/pipermail/r-help/2007-October/143931.html
print("There are no birthdays on")
## [1] "There are no birthdays on"
do.call("c", onlynulldates)
##  [1] "2014-02-08" "2014-02-21" "2014-02-22" "2014-03-06" "2014-04-16"
##  [6] "2014-04-21" "2014-05-03" "2014-05-24" "2014-06-26" "2014-08-03"
## [11] "2014-08-06" "2014-08-23" "2014-11-11" "2014-11-13" "2014-12-06"
## [16] "2014-12-13" "2014-12-23"