Insights and

Strategies: • An Analysis of Grocery Store

Transactions
• Teresa Rose B
Analyzing Grocery • 27th OCT 2024

Store Data
CONTENT
1. Exploratory Analysis of
data
2. Market Basket Analysis
3. Associations
Identification
4. Suggestion of Possible
Combos with Lucrative
Offers
 Agenda

EXECUTIVE SUMMARY EXPLORATORY DATA MARKET BASKET ASSOCIATIONS RECOMMENDATION

OF THE DATA ANALYSIS ANALYSIS IDENTIFIED
 Problem statement

Executive
summary Executive Summary &
of the Data Dictionary

data

Assumptions about data

Problem
statement
• A Grocery Store shared the
transactional data with you. Your
job is to identify the most
popular combos that can be
suggested to the Grocery Store
chain after a thorough analysis
of the most commonly occurring
sets of items in the customer
orders. The Store doesn’t have
any combo offers. Can you
suggest the best combos &
offers
Executive Summary

Data Overview:
➢ Timeframe: 01-01-2018 to 26-02-2020
➢ Dataset Size: 20,641 Rows, 3 Columns
➢ Missing Values: None
➢ Duplicate Values: 4,730
Objective:
➢ Conduct a thorough analysis of Point of Sale (POS) data to provide
recommendations for increasing revenue through popular combo offers and
discounts for customers.
Exploratory Analysis:
➢ Insights: The exploratory analysis reveals key trends and patterns in sales,
providing a comprehensive understanding of customer purchasing behaviors.
Market Basket Analysis:
➢ Methodology: Association rules were utilized to identify relationships between
products purchased together.
➢ Findings: The analysis highlighted frequently purchased items, enabling the
creation of lucrative offers aimed at enhancing customer engagement and driving
sales.
 Column Name Data Type Description

The date on which the

Date Date transaction occurred (format:
Data YYYY-MM-DD)
Dictionary A unique identifier for each
Order_id Integer
order placed by customers

The name of the product

Product String
purchased in the transaction
Assumptions:

• The data consists of a list of items purchased at a grocery store on

various dates.
• Each entry represents a single item purchased.
• The first column indicates the date of purchase.
• The second column identifies the customer who made the
purchase.
• The third column lists the item purchased.
• The same item may be purchased by multiple customers on
different dates.
• There is no information available regarding the quantity or price of
each item.
• Duplicated values have not been removed from the dataset.
Exploratory Analysis

• Exploratory Analysis of data (Trends

across months/years/quarters/days)
• Summary
Product
Frequency
Orders by
• Based on the available data, it can be inferred that there was a relatively
stable level of order activity from 2018 to 2019, with only a slight
decrease noted between these two years. However, a more detailed
analysis is necessary to understand the trends and patterns in order

Year
data for 2020. It’s important to consider that the dataset currently
includes only the first two months of 2020, which may not fully
represent the overall yearly trends.

• Further investigation is needed to draw comprehensive conclusions

about customer behaviour and order dynamics in 2020.
Orders by
• It can be inferred that there was a general
downward trend in the number of unique order
IDs from Q1 to Q3. This may indicate a potential

Quarterly decrease in overall order activity during this

period. However, further analysis is required to
understand the complete picture of order trends
throughout the year, especially in the absence of
Q4 data.
 • The data indicates variations in customer ordering patterns
Orders by Month throughout the year.
• The highest number of unique orders was observed in January,
followed by February and May. June had the lowest count of
unique orders. It is worth noting that there is a relatively
consistent level of order activity during the summer months
(June, July, and August) and a slight increase in September
 • The data indicates relatively consistent order activity
Orders by Week throughout the weekdays, with a slight increase in unique
orders observed on Wednesday, Thursday, and Friday. Notably,
Saturdays and Sundays show the highest count of unique
orders, suggesting increased customer engagement and
purchasing behaviour over the weekends.
• This trend highlights the importance of weekend promotions
and marketing strategies to capitalize on the higher customer
activity during this period.
 •

Top ordered
Cereals: 291 orders
• Poultry: 287 orders
• Flour: 280 orders

products- 2018 •
•
•
Shampoo: 275 orders
Lunch Meat: 269 orders
These products showcased significant popularity among customers in 2018,
with cereals leading the orders closely followed by poultry. The strong
demand for flour, shampoo, and lunch meat indicates a diverse range of
consumer preferences and highlights key items in the market during this
year.
 • Poultry: 288 orders

Top ordered • Soda: 287 orders

• Dishwashing Liquid: 271 orders

products- 2019 • Waffles: 270 orders

• Milk: 270 orders
• These products highlighted strong customer demand throughout 2019,
with poultry leading the way, closely followed by soda. Dishwashing
liquid, waffles, and milk also received substantial orders, reflecting
diverse consumer preferences.
 • Dinner Rolls: 67 orders

Top ordered • Poultry: 65 orders

• Pork: 63 orders

products- 2020 • Ice Cream: 58 orders

• Beef: 58 orders
• These products demonstrated significant customer demand in 2020,
with dinner rolls leading the pack, followed closely by poultry. Both
pork and ice cream were also popular choices, showcasing a diverse
range of preferences among customers.
Summary
• S ale s I ns ights

• T op-S el ling Products:

• 2018 & 2019: Cereals , Poultry , a nd S oda were the top three produc ts sold.

• Ja nua ry & F ebrua ry 2020: Dinne r roll s, Poultry, a nd Pork toppe d the sa le s.

• Consistent Hig h Sell ers:

• Poultry , Ce rea ls, a nd S oda ha ve bee n consiste ntly hig h-se lling products ov er the ye ars .

• N on-F ood Produc ts :

• T he hig hes t-selling non-food products a re S oap a nd T oilet Pa per, whil e Ha nd Soa p ra nks as the lowest seller.

• E ata bles Ca teg ory:

• T he top-se lling products in the e ata bles ca te g ory are Poul try , S oda , a nd Ce re als , wi th Pork, F ruits, and S andwich Loa ve s
be ing the le as t sold.

• S al es Trends by D ay:

• M os t products were s ol d on S unda ys , whereas Monday s rec orded the leas t sales.

• M onthly S al es Trends :

• 2018: Hig hest s ale s occ urre d in January ; lowe st in F ebruary.

• 2019: Ma rch sa w the hig he st sa le s, while January ha d the lowe st.

• Q uarterl y Performa nce:

• T he hig hes t sales were recorded in Q1 2019 a nd Q 3 2018.

• Q 2 Sa les Consistency:

• T he c ount of products sold in Q 2 was a pproxim ate ly the sa me for 2018 a nd 2019.

• 2020 Sa les D ata Li mita tion:

• S ale s c ount for 2020 is l ow , lik e ly due to da ta be ing ava ila ble only unti l Fe bruary 26.
Recommendations
Promote a nd Stock T op -S el ling Products:

• F oc us on poultry, soda, and ce re als as the y c onsi stently ra nk a mong the top-se lling products.

I nc rea se Stock of H ig h-Dema nd Non-F ood I tem s:

• Consider boosting inventory for soap a nd toi let pa per, whic h are the hig hes t-selling non-food products.

I nv estiga te Low S al es of H and S oa p:

• E va luate fa ctors contributi ng to the l ow s ale s of hand soa p and imple me nt strate gie s to e nhance its s ale s.

S chedule Promotions on Sunday s:

• Pla n promotions a nd spec ial offers on S unda ys to c apita lize on the hig hest s ale s da y of the we ek .

Targ et L ow Sa les M onth :

• De ve lop ma rke ting c am pa ig ns and disc ounts during F e brua ry to addres s historica lly l ow s ale s during tha t month.

Ca pi tal ize on Hig h Sa les M onths :

• Pla n mark eting initia ti ve s and disc ounts for January and Ma rch to boos t sale s during the se historica lly hig h -sa le s months.

Replica te Successful S a les Patterns :

• Aim to replic ate the s uc cessful sa les patterns observed in Q 1 2019 a nd Q3 2018.

M ai nta in Consis tent Stock Levels :

• E nsure that s toc k le ve ls for Q 2 products a re c ons istent with prev ious ye ars to s us ta in sa le s performanc e.

Consider Lim ited 2020 D ata:

• Keep in m ind the li mited da ta a va ila bl e for 2020 when ma ki ng sa les and m arketing deci sions.
Market
• Market Basket Analysis Meaning
• MRA KNIME Work Flow & Output Table

Basket
Analysis
Market
• Market Basket Analysis is a technique used to identify patterns and
associations among products that are frequently purchased together in
customer transactions. It helps businesses understand customer behaviour
and uncover product relationships, which can be used to optimize sales
strategies and increase revenue. By analysing transactional data, Market

Basket Basket Analysis generates insights on item co-occurrence and association

rules, enabling businesses to make informed decisions on product bundling,
cross-selling, and targeted marketing campaigns. This analysis provides
valuable insights into customer preferences, allowing businesses to improve

Analysis
customer satisfaction and drive business growth.
Association
• Association rules in Market Basket Analysis reveal the
relationships and co-occurrence patterns between items,
providing valuable insights into customer purchasing behaviour

Rules & its

and preferences.
• The relevance of association rules lies in their ability to guide
businesses in optimizing product placement, creating targeted
marketing campaigns, and implementing effective cross-selling

relevance. and upselling strategies to enhance customer satisfaction and

increase revenue.
KNIME workflow
Threshold values
• The threshold value for Minimum Support is set at 0.05.
• The threshold value for Minimum Confidence is set at 0.6.

of Support and
• These threshold values are used to identify significant and trustworthy association rules.
• Minimum Support helps filter out rules with insufficient support in the dataset.

Confidence
• Minimum Confidence ensures that only rules with a high likelihood of occurrence are
considered.
• By applying these thresholds, we can focus on rules with strong support and confidence,
leading to meaningful insights for decision-making.
 • Present the associations in a table format.
Associations • Explain the calculated values of support, confidence,
and lift.
Identification
 Support Confidence Lift Recommended_item Recommended_with Items_list
0.050043898 0.640449438 1.700400723 juice <--- [yogurt, toi let paper, alumi num foil]
0.050043898 0.619565217 1.644952873 juice <--- [yogurt, poultry, aluminum foil]
0.050043898 0.612903226 1.615964755 coffee/tea <--- [yogurt, cheeses, cereals]
0.050043898 0.6 1.42375poultry <--- [dishwashing liquid/detergent, l aundry detergent, mixes]
0.050921861 0.630434783 1.677722471 mixes <--- [yogurt, poultry, aluminum foil]
0.050921861 0.610526316 1.659640749 sandwich bags <--- [cheeses, bagels, cereal s]
0.050921861 0.674418605 1.726208518 cheeses <--- [bagels, cereals, sandwich bags]
0.050921861 0.617021277 1.55828655cereals <--- [cheeses, bagels, sandwich bags]
0.050921861 0.630434783 1.620914712 dinner roll s <--- [spaghetti sauce, poult ry, cereals]
0.050921861 0.637362637 1.512408425 poultry <--- [dinner rolls, spaghet ti sauce, cereal s]
0.050921861 0.604166667 1.589251347 milk <--- [poultry, laundry detergent, cereals]
0.051799824 0.627659574 1.610144719 eggs <--- [dinner rolls, poult ry, soda]
0.051799824 0.641304348 1.648861517 dinner roll s <--- [spaghetti sauce, poult ry, ice cream]
0.051799824 0.686046512 1.627931202 poultry <--- [dinner rolls, spaghet ti sauce, ice cream]
0.051799824 0.627659574 1.613779357 dinner roll s <--- [spaghetti sauce, poult ry, juice]
0.051799824 0.602040816 1.428592687 poultry <--- [dinner rolls, spaghet ti sauce, juice]
0.051799824 0.634408602 1.627458103 eggs <--- [paper towels, dinner roll s, pasta]
0.051799824 0.602040816 1.621098085 pasta <--- [paper towels, eggs, di nner rolls]
0.053555751 0.642105263 1.650920756 dinner roll s <--- [spaghetti sauce, poult ry, laundry detergent]
0.053555751 0.655913978 1.556429211 poultry <--- [dinner rolls, spaghet ti sauce, laundry detergent]
0.055311677 0.623762376 1.564901644 ice cream <--- [paper towels, eggs, pasta]
0.055311677 0.63 1.616148649 eggs <--- [paper towels, ice cream, pasta]
0.055311677 0.642857143 1.73100304pasta <--- [paper towels, eggs, ice cream]
0.055311677 0.649484536 1.79119343paper t owels <--- [eggs, i ce cream, pasta]

Association
represented in
tabular form
Market basket analysis

• Support:
• Represents the probability of observing items together in a transaction.
• Calculated as the number of transactions containing both items divided by the total number of
transactions.
• Measures the frequency of the itemset in the dataset.
• High support indicates popularity, suggesting the itemset should be considered for promotion or joint
placement.
• Confidence:
• Represents the conditional probability that a transaction containing one item also contains another
item.
• Calculated as the number of transactions containing both items divided by the number of transactions
containing the first item.
• Measures the strength of the association between two items.
• High confidence suggests that the items are likely to be purchased together, making them suitable for
recommendations.
• Lift:
• Measures how much more often two items occur together than expected if they were independent.
• Calculated as the support of the itemset divided by the product of the individual supports of the items.
• A lift value of 1 indicates independence, while a value greater than 1 indicates a positive association.
• A lift value less than 1 suggests a negative association.
• High lift values indicate a strong association, useful for cross-selling or bundling strategies.
• 4o mini
Suggestion of
Possible Combos
with Lucrative
Offers
• Write recommendations
• Make discount offers or combos (or buy two get
one free) based on the associations and your
experience
Recommendations
(discount offers/
combos)

• Combo Deal: Offer a special combo deal where customers can buy yogurt, poultry, and aluminium foil along with
juice to avail a discounted price or additional item.
• Buy Two Get One Free: Introduce a "buy two get one free" offer on dinner rolls, spaghetti sauce, and ice cream to
incentivize customers to purchase these items together.
• Bundle Promotion: Create a bundle promotion where customers can buy paper towels, eggs, and pasta together at a
discounted.
• Cross-Selling Offer: Provide a cross-selling offer where customers purchasing cereals can get a discount on cheese,
bagels, and sandwich bags.
• Limited-Time Promotion: Launch a limited-time promotion where customers buying poultry, laundry detergent, and
mixes can receive a percentage savings.
• Loyalty Program: Implement a loyalty program where customers who frequently purchase recommended items or
participate in the suggested combos can earn rewards or exclusive discounts. This will incentivize customer loyalty
and encourage them to continue shopping with the store, fostering long-term relationships and repeat purchases.
• These recommendations are based on the association rules and the occurrence of certain items together, aiming to
increase customer satisfaction and encourage them to explore additional products.
Thank you