Projects

Grocery Sales Analysis

Initial exploratory analysis of Kaggle’s Supermart Grocery Sales – Retail Analytics Dataset.

View the public sales findings report · Open the report file.

The source CSV is in data/. The dataset contains order-level grocery sales, including order date, category, subcategory, city, region, sales, discount, and profit. The data represents Tamil Nadu, India, and covers 2015–2018.

Run

.venv/bin/python script/sales_analysis.py

The script writes summary tables, a text summary, and four charts to outputs/. Hyphenated order dates are interpreted as day-month-year; slash-formatted dates are parsed as month/day/year. The parser reports any failures in outputs/analysis_summary.txt.

Build the standalone HTML report after running the analysis:

.venv/bin/python script/build_html_report.py

This writes sales_findings.html with the conclusion first, a clearly attributed data source, 10 source sample rows, the cleaning procedure, and the analysis. Charts are embedded, so the report can be opened or shared as a single file.

The HTML report is self-contained and needs no companion assets. Generated analysis files in outputs/ and the duplicate Kaggle ZIP are excluded from version control. The source CSV is kept in data/ so the analysis can be reproduced; the report itself can be viewed without it. The proposed fact/dimension design and its source-data limitations are documented in data/data_model.md.

SQL analysis

The SQLite CLI script imports the same source CSV and writes the monthly, category, and region summaries plus annual sales, discount-band, and data quality CSVs to outputs/. From this project folder, run:

mkdir -p outputs
sqlite3 outputs/sales_analysis.db < script/sales_analysis.sql

The SQL script is independent of the Python scripts and does not build the HTML report. Install SQLite if the sqlite3 command is unavailable.

Initial questions

• How do monthly sales vary over time?
• Which product categories and regions contribute most to sales and profit?
• How do discount levels relate to profit and profit margin?
• Which cities and subcategories are strongest or weakest?

The dataset is suitable for exploratory analysis, but its Kaggle description should be consulted before using findings for real operational decisions.