Challenge
As an analyst for Instacart, an online grocery stpre that operates through an app. Instacart already has very good sales, but they want to uncover more information about their sales patterns. My task is to perform an initial data and exploratory analysis of some of the company’s data in order to derive insights and suggest strategies for better segmentation based on the provided criteria.
Context
The Instacart stakeholders are more interested in the variety of customers in their database along with their purchasing behaviors. They assume they can’t target everyone using the same methods, and they’re considering a targeted marketing strategy. They want to target different customers with applicable marketing campaigns to see whether they have an effect on the sale of their products.
Project Scale
3 weeks
Data
- “The Instacart Online Grocery Shopping Dataset 2017" Accessed from www.instacart.com/datasets/grocery-shopping-2017
- Data Dictionary https://gist.github.com/jeremystan/c3b39d947d9b88b3ccff3147dbcf6c6b
Primary Stakeholder
CareerFoundry Data Analytics Course
Skills
- Python
- Data Wrangling and merging
- Deriving variables
- Grouping datasets
- Aggregating data
- Reporting in Excel
Questions from the Sales and Marketing Team
What are the busiest days of the week and busiest hours of the day?
Days and hours where most money spent?
Are certain types of products more popular than others?
Analyze different types of customers based on...
Their loyalty
Their region
Their age
The Process
To illustrate the data cleaning and merging process, I used a population flow chart. The steps were as follows:
- Cleaned the “orders” dataset
- Cleaned the “orders_products_prior” dataset
- Transformed the “products” dataset
- Cleaned the “customers” dataset
- Merged all datasets into a final dataset
- Performed column derivations and aggregations to address Sales and Marketing questions
Customer Habits Analysis
Saturday is the busiest day for orders.
With the slowest days being Tuesday and Wednesday
The majority of orders are between 9am-4pm.
With the slowest times being between midnight and 6am.
Most popular orders are Produce and Dairy Eggs.
With the least popular being Other and Bulk items.
I defined Loyal Customer as someone who orders on the app over 40 times a month, Regular Customer as someone who orders less than 40 but more than 10 times, and a New Customer as someone who orders under 10 times.
There are more Low-Spenders than High-Spenders.
High-Spenders are people who spend over $15 per order on average, while Low-Spenders are people who spend under that on average.
Since the app requires you to be over 18 to use it, Young Adult classifies people from the age of 18-35, Middle Aged is people between 36-64, and Senior is anyone over the age of 65.
Recommendations
- Weekend Sales: Since people tend to spend more on weekends, try offering incentives like discounted delivery fees or a points system to encourage purchases during weekdays and off-peak hours.
- Customer Loyalty: Implement a rewards program to turn regular customers into loyal ones, encouraging more frequent visits.
- Product Focus: Keep popular items like produce and dairy, and consider adding new vegetarian options to cater to demand and potential preference for fresher, shorter-shelf-life products.
- Regional Marketing: Focus marketing efforts on regions with fewer customers and feature a “local favorite” section to boost interest and sales.
- Targeted Advertising: Direct your advertising towards middle-aged adults, as they make up the bulk of Instacart’s customer base.
Retrospective
What went well?
This project was a valuable experience with the most rewarding aspect being the strategizing for the analysis and discovering which techniques would yield the best results.
What didn’t go well?
With limited prior experience in Python, I faced some challenges. Creating derived columns required careful thought, and I occasionally revisited earlier steps to ensure accuracy.
Despite these hurdles, once I became more comfortable with the tools, the project flowed smoothly.
Future steps
The next step would be to conduct a detailed analysis of price distribution throughout the day to gain a deeper understanding of purchasing behaviors.
Final thoughts
This was my first python project. It was great to experience how to use python in a way that could create datasets to explain my analysis findings.
Want to see more?
Check out my Github to see more recommendations and my Python scripts.