Statement of the Problem

General Statement of the Problem

How does the market basket of consumers in Bulacan be analyzed?

Statement of the Problem

Problems

1. What is the demographic profile of the respondents?
2. How may the degree of importance of buying factors in making decisions was described?
3. What is the frequency of visits of customers to stores?
4. What are the top ten grossing group of products that are associated together in one basket or cart?
5. For whom are the products purchased?
6. What is the most preferred method of purchase?
7. What is the most effective sales promotion device according to the consumers?

Significance of the Study

Who will Benefit?

1. Supermarket or Grocery Stores
2. Store Managers
3. Producers and Suppliers
4. Consumers / Customers
5. Marketers and Sales People
6. Future Researchers

Scope and Delimitation

The study is targeted to general consumers which the population is Bulacan. The respondents are not constrained with the popular stores nor with the small players. To avoid bias, the research is mostly conducted at food chains not within the store.

Excluded are the POS transactions from the cashiers.

The survey is targeted on basic necessities, prime commodities and Department of Trade and Industry (DTI) list which were merged to form a selection of items.

Theories

Market Basket Analysis - Frequency, Support, and Confidence.

Frequency is the number of times the items are bought together with the target product or item. This shows the most related item to the target item. Support on the other hand shows which pairs are purchased a lot that supports relation to the target item, although low support doesn’t mean that the item will not be ordered together. The confidence is the most important one as this shows the relationship of support and frequency. But the real interesting item to know are those with support and confidence of both high or low values, which gives importance to lift, these are the interesting values that supports association.

Theories

Affinity Analysis.

A type of analysis which you are looking for items that go together, especially those in the purchased transactions. Data recorded are analyzed on how they are related among other data. This analysis is part of data mining applied into retail business which is known as market basket analysis. Other application includes fraud examination and detection for those that will deviate from the rule sets inferred from the model.

Theories

Recommendation Theory.

Predicting the item that is most relevant to the previous choices is the goal of recommendation. Hardesty (2014) reviewed the group at MIT’s Laboratory for Information and Decision Systems (LIDS). LIDS specializes in analyzing how social networks process information. They are doing predictions, but they used collaborative filtering as their algorithm. Their theory is not the typical collaborative filter, wherein they just assign the likelihood of choice among different samples. They produce sorted probability clusters as a new approach, and also those that do not select an item are assigned a cluster, otherwise choice so it would affect the correlation of choices of the customers. This approach is more principled according to Sujay Sanghavi, a professor of electrical and computer engineering at the University of Texas, Austin. Recommendation is also applied to services that Netflix and Amazon. Pennock, et al. (2000), have studied the collaborative filtering (CF), which have been greatly used in the internet as confirmed by Microsoft Research, also they’ve found out that it is hard to measure performance of CF.

Conceptual Framework

Input-Process-Output Chart

Methods and Techniques of the Study

The study is a quantitative study targeted at consumers in Bulacan. The use of questionnaire with choices and use of scales are incorporated with the questionnaire. Most market basket analysis studies used POS transactions, while this study uses the data answered through questionnaire. This approach removes bias to brands.

Population of the Study

Respondents are from Bulacan with a population of 3,292,000 people. For a sample to be valid it should be less than ten percent (10%) of the population and at least 30 respondents. A modified slovin formula was used for the computation. Selected towns for the study are Malolos, Baliwag, Bustos, Plaridel and Pulilan, where is care is taken as using proper proportion per town.

Research Instruments

Survey-questionnaire method is used in this study. Use of R Statistical Language was used in the data analysis and presentation of the data.

Collection and Data Gathering Procedure

Survey was performed by using random sampling, while maintaining the proper proportion of respondent per town which is based on the population per town. An interviewer asked questions and recorded the answer on the paper questionnaire. Data is then encoded into the computer by using Google Forms to simplify the encoding.

Data Processing and Statistical Treatment

Data processing was done using an open-source software R. R is a popular statistical programming language used by international schools and universities in Australia, New Zealand and US. To speed up the computations, use of the likert package for R was used, ggplot2 package for visualization, flexdashboard package for presentation and data storytelling, and shiny package for the dashboard application. The process used for statistical treatment is the standard data analysis, however best practices for data science and analytics was used to produce correct output and make sure that the steps can be reproduced in an instant even at the time of this presentation. RStudio is a graphical user interface aids in debugging the statistical program code used for analysis.

Statement of the Problem

Problem 1

1. What is the profile of the respondents in terms of the following:

• 1.1 civil status;
• 1.2 age;
• 1.3 gender;
• 1.4 monthly net income; and
• 1.5 highest educational attainment

 [1] "Timestamp"                                   
 [2] "Name..Optional."                             
 [3] "Location"                                    
 [4] "Civil.Status."                               
 [5] "Age"                                         
 [6] "Gender"                                      
 [7] "Monthly.Net.Income"                          
 [8] "Highest.Educational.Attainment"              
 [9] "Public.Transportation"                       
[10] "Proximity"                                   
[11] "Parking.Space"                               
[12] "Cheap.Prices"                                
[13] "Grocery.Store"                               
[14] "Sari.sari.Store"                             
[15] "Mall...Hypermarket"                          
[16] "Supermarket"                                 
[17] "Wet.Market"                                  
[18] "Dry.Market"                                  
[19] "Items"                                       
[20] "Who.are.the.consumers.of.purchased.products."
[21] "Most.preferred.Method.of.Payment"            
[22] "Most.effective.Sales.Device.for.You"         

'data.frame':   328 obs. of  22 variables:
 $ Timestamp                                   : Factor w/ 328 levels "2017/07/16 11:51:23 AM GMT+8",..: 1 2 3 4 5 6 7 8 9 10 ...
 $ Name..Optional.                             : Factor w/ 315 levels "---","----","-----",..: 296 244 236 247 291 259 270 286 127 277 ...
 $ Location                                    : Factor w/ 5 levels "Baliwag","Bustos",..: 2 2 2 2 2 2 2 2 2 2 ...
 $ Civil.Status.                               : Factor w/ 4 levels "Married","Separated",..: 1 1 3 1 3 3 2 4 3 1 ...
 $ Age                                         : int  63 62 65 78 22 22 54 62 28 50 ...
 $ Gender                                      : Factor w/ 2 levels "Female","Male": 1 1 2 2 1 1 1 1 1 2 ...
 $ Monthly.Net.Income                          : int  2000 20000 6000 5000 9000 8000 5000 15000 10000 2000 ...
 $ Highest.Educational.Attainment              : Factor w/ 4 levels "College","Elementary",..: 2 1 1 3 1 1 3 1 1 1 ...
 $ Public.Transportation                       : int  2 3 2 2 1 2 3 2 1 2 ...
 $ Proximity                                   : int  2 3 2 2 2 2 5 3 1 2 ...
 $ Parking.Space                               : int  3 3 2 1 1 2 4 2 1 2 ...
 $ Cheap.Prices                                : int  1 1 1 1 2 3 2 1 1 1 ...
 $ Grocery.Store                               : int  3 2 1 1 1 1 2 2 3 2 ...
 $ Sari.sari.Store                             : int  2 2 1 1 1 1 2 2 2 1 ...
 $ Mall...Hypermarket                          : int  3 3 1 2 4 4 4 2 2 4 ...
 $ Supermarket                                 : int  3 1 2 2 4 3 3 3 3 4 ...
 $ Wet.Market                                  : int  2 1 1 2 4 4 2 3 3 1 ...
 $ Dry.Market                                  : int  2 1 1 2 4 4 2 2 3 1 ...
 $ Items                                       : Factor w/ 322 levels "Bread in any shape and name; excluding pastries and cakes;Coffee;Sugar;Powdered; liquid; bar; laundry and detergent soap;Candie"| __truncated__,..: 152 308 293 58 317 123 37 101 200 181 ...
 $ Who.are.the.consumers.of.purchased.products.: Factor w/ 25 levels "Children","Children;Parents;Relatives",..: 19 17 5 16 23 23 15 3 10 15 ...
 $ Most.preferred.Method.of.Payment            : Factor w/ 3 levels "Cash","Credit Card",..: 1 1 1 1 1 1 1 1 2 1 ...
 $ Most.effective.Sales.Device.for.You         : Factor w/ 4 levels "Demonstratrion (free taste, free sample, etc.)",..: 3 2 3 3 3 3 3 4 3 3 ...
NULL

• Refactoring was done and use of a midpoint function was for the computation of intervals.

[1] "Location"                       "Civil.Status."                 
[3] "Age"                            "Gender"                        
[5] "Monthly.Net.Income"             "Highest.Educational.Attainment"

• Scales factor are prepared by using the refactor variable and also the use of the likert package.

Statement of the Problem

Problem 2

2. How may the degree of importance of buying factors in making decisions be described in terms of:

• 2.1 availability of public transportation;
• 2.2 proximity of the store;
• 2.3 ample parking space; and
• 2.4 cheap prices?

                                                      
transpo   HI: 47   CI:170   SI:100   LI: 11   NI:  0  
proximity HI: 43   CI: 77   SI: 97   LI:101   NI: 10  
parking   HI: 19   CI: 70   SI:121   LI: 81   NI: 37  
cheap     HI:245   CI: 80   SI:  3   LI:  0   NI:  0  

Item	HI	CI	SI	LI	NI
transpo	14.3292682926829258	51.829268292682926	30.48780487804878092	3.3536585365853662	0.0000000000000000
proximity	13.1097560975609753	23.475609756097558	29.57317073170731447	30.7926829268292686	3.0487804878048781
parking	5.7926829268292686	21.341463414634145	36.89024390243902474	24.6951219512195124	11.2804878048780495
cheap	74.6951219512195053	24.390243902439025	0.91463414634146334	0.0000000000000000	0.0000000000000000

Statement of the Problem

Problem 3

3. What is the frequency of visits of customers to stores?

• 3.1 grocery store;
• 3.2 sari-sari store;
• 3.3 mall/hypermarket;
• 3.4 supermarket;
• 3.5 wet market; and
• 3.6 dry market

                                                                
 Grocery         AT   : 43   MT   :138   NO/NS:115   ST   : 31  
Sari-stari Store AT   : 69   MT   :157   NO/NS: 87   ST   : 14  
Hypermarket      AT   : 10   MT   : 71   NO/NS:125   ST   :120  
Supermarket      AT   : 18   MT   : 67   NO/NS:122   ST   :119  
Wet Market       AT   : 19   MT   :100   NO/NS:128   ST   : 77  
Dry Market       AT   : 21   MT   : 97   NO/NS:126   ST   : 81  
                            
 Grocery         NA   :  1  
Sari-stari Store NA   :  1  
Hypermarket      NA   :  2  
Supermarket      NA   :  2  
Wet Market       NA   :  4  
Dry Market       NA   :  3  

Item	AT	MT	NO/NS	ST	NA
Grocery	13.1097560975609753	42.073170731707314	35.060975609756099	9.4512195121951219	0.30487804878048780
Sari-stari Store	21.0365853658536572	47.865853658536587	26.524390243902442	4.2682926829268295	0.30487804878048780
Hypermarket	3.0487804878048781	21.646341463414632	38.109756097560975	36.5853658536585371	0.60975609756097560
Supermarket	5.4878048780487809	20.426829268292682	37.195121951219512	36.2804878048780495	0.60975609756097560
Wet Market	5.7926829268292686	30.487804878048781	39.024390243902438	23.4756097560975583	1.21951219512195119
Dry Market	6.4024390243902438	29.573170731707314	38.414634146341463	24.6951219512195124	0.91463414634146334

Statement of the Problem

Problem 4

4. What are the top ten grossing group of products that are associated together in one basket or cart?

• 4.1 Using Market Basket Analysis - top products that goes together in terms of


• 4.1.1 frequency;
• 4.1.2 support;
• 4.1.3 confidence; and
• 4.1.4 lift.

Statement of the Problem

Problem 4

4. What are the top ten grossing group of products that are associated together in one basket or cart?

• 4.2 What are the products that goes together?

Statement of the Problem

Problem 5

5. For whom are the products purchased?

• 5.1 Own Self
• 5.2 Partner (Wife / Husband)
• 5.3 Children
• 5.4 Parents
• 5.5 Relatives
• 5.6 Senior Citizen (age 60 and above)
• 5.7 Sick members of the immediate family
• 5.8 Friends
• 5.9 Charity

Statement of the Problem

Problem 6

6. What is the most preferred method of purchase?

Statement of the Problem

Problem 7

7. What is the most effective sales device according to the consumers?

Summary of Findings

1. What is the profile of the respondents in terms of the following: The profile of the respondents is taken from the five towns. Majority of the respondents are from Malolos (37.5%), followed by Baliwag (22.56%), Plaridel (16.16%), and Bustos (10.06%). The total number of respondents were 328 from a population of 3,292,000 in Bulacan. Most of the respondents are married (59.76%) and the least number of respondents are separated (4.27%). Youngest respondent was 19 years old, and the oldest was 81 years old. The mean age is 39 years old and the median age is 36 years old. There are 178 female respondents (54.27%) and 150 male respondents (45.73%). The median monthly net income in Bulacan is 6,000.00 pesos and the average monthly net income is 7,731.26 pesos.
   Among the respondents’ educational attainment are as follows: majority has completed High School (457.73%), followed by College degree (44.21%), Elementary education (9.76%) and Vocational (0.30%). It seems that from the sample most of the respondents haven’t had a graduate degree, it may have happened as the survey was conducted during working hours and the possible respondent with a graduate degree was teaching in an educational institution or working in large firms.
2. How may the degree of importance of buying factors in making decisions be described?
   Looking into the buying factors, in terms of public transportation, proximity of the store, ample parking space and cheap prices are taken as follows: most respondents have considered that cheap prices are highly important and the least is parking.
3. What is the frequency of visits of customers to stores? Consumers have different frequency of visits to different store types. The following store types are included in the questionnaire: grocery store, sari-sari store, hypermarket, supermarket, wet market, and dry market. Most people prefers to purchase from sari-sari store and the least at hypermarket.
4. What are the top ten grossing group of products that are associated together in one basket or cart? (This is visible from the application created). The following are the top 10 group of products.

1. {Chili,Cooking Oil,Garlic} to {Pepper}
2. {Chili,Cooking Oil,Garlic,Onions} to {Pepper}
3. {Chili,Cooking Oil,Garlic,Rice} to {Pepper}
4. {Chili,Cooking Oil,Garlic,Onions,Rice} to {Pepper}
5. {Bath Shampoo,Coffee,Fresh Pork / Beef,Sugar} to {Coffee creamer}
6. {Bath Shampoo,Coffee,Fresh Pork / Beef,Rice,Sugar} to {Coffee creamer}
7. {Chili,Cooking Oil,Onions} to {Pepper}
8. {Chili,Cooking Oil,Onions,Rice} to {Pepper}
9. {Bath Shampoo,Coffee,Cooking Oil,Fresh Pork / Beef} to {Coffee creamer}
10. {Bath Shampoo,Coffee,Cooking Oil,Fresh Pork / Beef,Powdered,liquid,bar,laundry and detergent soap} to {Coffee creamer} Details are in the output application.

5. For whom are the products purchased? Consumers purchase products for their own self (36.38%), followed by for children (24.75%), partner (wife/husband) (22.71%), parents (10.62%), relatives (4.29%), senior citizen (1.13%), and friends (0.11%).
6. What is the most preferred method of purchase? Most preferred method of purchase in Bulacan shows that most prefers cash method (99.09%), followed by credit card (0.61%), and debit card (0.30%).
7. What is the most effective sales promotion device according to the consumers? The most effective sales device in Bulacan is pricing (57.93%), followed by demonstration (23.17%), display (9.76%), and sales talk (9.15%). The top ten products purchased in Bulacan are: rice, fresh vegetables / root crops, powdered liquid bar (laundry and detergent soap, poultry meat (like chicken), fresh eggs, cooking oil, fresh fish (and canned fish, and dried fish), fresh pork / beef, instant noodles, and coffee.

Conclusions

1. Balanced on gender response, most respondents are high school, followed by college and least are with vocational courses. Monthly net income’s median is 6,000.00 pesos.
2. Cheap prices are mostly preferred by people in Bulacan, which they prefer to save more of what they earn.
3. Most people visits sari-sari store which is still more common and more convenient for people in Bulacan to reach.
4. People buy products that includes pepper and coffee creamer.
5. People buy products for themselves, children and partner, and least for their friends.
6. Using cash is the most preferred method in Bulacan. Very few use credit cards and debit cards, due to risk of using it and trust is still in the province.
7. Nothing beats the discounts, pricing on the consumer preferences, next is the product demonstrations, while the display and sales talk were considered low by the consumers in Bulacan, people still saves most of their money.

Recommendation

1. Managers should consider the insights given by the data and numbers not just based on experience and gut feeling. Data cannot lie and can help you see the real picture of the consumer preferences.
2. Creating promotions of which items to bundle together will be easier with the use of market basket analysis. This will provide the better sales since this would include the commodities that consumers will really buy together.
3. Improvement on store layouts of groceries, hypermarket and the like could be using the itemset in optimizing the store layout. You can now optimize the placements of commodities, you can decide where to place the items which will result in easier shopping for the consumers.
4. Creation of intelligent systems for other businesses such as for a convenience store which analyses the current items listed being punched can help the cashier offer the customer of another product, since the use of the association rules you can recommend another item based on the chosen items by the customer checking out at a convenience store.
5. For the fast food industry, understanding the individual items bought together can help create a new value meal which contains the most purchased items.
6. It is good to have programs that will help consumers use other methods of payment such as the use of debit card and credit cards for the consumer’s convenience.
7. For future researchers, if they will reproduce a study on market basket analysis, it is better to create a dashboard on a data product application to simplify the communication of results.

Thank you.
Rowen Remis R. Iral
http://wenup.wordpress.com