Instacart Basket Analysis
Python Project
conducted April 2024
Contents:
Goals
Tools & Skills Used
Overview
Recommendations
Goal:
This project was an assignment from my data analytics course.
Instacart, an online grocery store, has great sales in the United States. They plan to uncover more info about their sales patterns.
My task is to analyze the sales data in order to derive insights and suggest strategies to further improve profitability.
Sources:
Tools:
Python
Excel
Skills Used:
Data wrangling
Data merging
Deriving variables
Grouping data
Aggregating data
Reporting in Excel
Population flows
Overview
Key Demographics:
The assignment asks us to identify the different types of customers and what they order. By identifying key demographics, we can find patterns and focus on the needs of our target audience to maximize sales.
In Python, I created an age profile flags and then converted them into percentages:
The largest demographic, Middle Aged users, make up of 40% of all users.
To find what customers were ordering, I created a bar chart of the orders dataset and grouped the age profiles to view what age group was ordering which product:
Top 5 purchased items are:
Produce
Dairy & eggs
Snacks
Beverages
Frozen
We can conclude majority of customers are ordering food for their family members.
Family Status:
The assignment next wanted us to analyze the family status of customers. Similarly, I used the same bar chart and grouped customers by single, married, living with parents & siblings, and divorced/widowed.
Looking at family status, the most prominent users were married customers. This confirms that these customers are ordering food for their family.
Peak Order Times:
After looking at the different types of customers and what they buy, we can analyze when orders are typically made.
The first bar chart shows the number of orders made each day. We can see that orders are high on Fridays, Saturdays, and Sundays.
The histogram below displays at what hour groceries are purchased. By 9AM, number of orders peak and plateau until around 4PM, then start tapering down throughout the night.
Recommendations
Advertisements:
Instacart is typically used on smartphones. To improve profitability, ad notifications can be pushed often during the slow days of the week (Monday through Thursdays) to remind customers to purchase groceries before the weekend rush or risk items going out of stock.
Key Demographics:
Considering most users are married, middle-aged customers with families, Instacart could cater to these users by offering coupons for buying products in bulk.
Users who are single tend to buy similar items as married customers, but at a smaller scale. These users could be offered similar coupons for purchasing the same products multiple times in the past.
Most & Least Ordered Products:
When on the checkout screen, we could advertise non-popular items that would compliment what the user is already purchasing. For example, if the customer were planing to buy fruits and frozen pizza, the app could suggest also buying meat/seafood, which is not ordered as often. Or suggesting alcohol when the customer orders snacks and beverages.
Limitations:
The data used was provided by the CareerFoundry data analytics course using fictional user data. Although the data cannot be used for real company decision making, it was good for practicing real world scenarios.
Next Steps:
I am interested in finding out what specific items are ordered in different regions of the country to find any trends. I’d also like to further analyze the smaller demographics and what products they purchase or leave in their cart.