Instacart is an online grocery store aiming to uncover deeper patterns in their sales data. This project involved a deep dive into over 32 million records to segment users and identify key drivers of sales across different US regions and demographics.
- Language: Python 3.x
- Key Libraries:
Pandas,NumPy,Matplotlib,Seaborn,Scipy. - Methodology: Professional Data Pipeline (Wrangling → Consistency → Profiling → Visualization).
- PII Redaction: Removed sensitive customer information (last names) to adhere to data privacy standards.
- Consistency Audits: Conducted rigorous checks for missing values, duplicates, and data type integrity.
- Combining Data: Merged multiple large-scale datasets (Orders, Products, Departments, Customers) into a unified analytical dataframe.
Developed new business-centric variables to drive deeper insights:
- Deriving Variables: Created flags for loyalty (New/Regular/Loyal), spending (High/Low spender), and order frequency.
- Grouping & Aggregating: Performed complex aggregations (mean, min, max) to compare spending habits across different customer profiles.
- Geographic Profiling: Grouped states into 4 major US Regions to analyze regional market performance.
- Large Scale Handling: Processed datasets exceeding 32,434,212 rows.
- Sampling Strategy: Implemented a 70/30 data split to execute complex visualizations while managing system memory constraints.
- Segmented Analysis: Correlated age, income, and family status with purchasing behavior.
- Department Performance: Identified the most popular product categories by region and time of day.
- Strategic Reporting: Translated technical findings into a final Excel report for executive stakeholders.
*Note: This project was developed as part of a professional Data Analytics certification by CareerFoundry.