Store Sales Analysis (Rossmann / Store Data)

Project for exploring store sales data, training forecasting models, and running exploratory notebooks and a small app.

Project structure

• app.py - main application entry (small demo / analysis runner).
• train.csv, store.csv - core datasets used for modeling and analysis.
• rossman.csv, rossman.ipynb, rossmann.ipynb - alternative dataset and notebooks.
• requirements.txt - Python dependencies for this project.

See the repository root for all files.

Overview

This repository contains preprocessing, exploratory analysis, and modeling artifacts for a store-sales forecasting task (Rossmann-style dataset). It includes Jupyter notebooks for interactive exploration and an app.py script for quick demos.

Requirements

• Python 3.8+ recommended
• A Conda environment data_scientist is used in this workspace (optional but recommended)

Install dependencies:

conda activate data_scientist
pip install -r requirements.txt

If you prefer venv:

python -m venv .venv
source .venv/Scripts/activate      # Windows: .venv\\Scripts\\activate
pip install -r requirements.txt

Datasets

• train.csv — primary training data (historical sales per store/day).
• store.csv — store meta-data (store type, assortment, competition info).
• rossman.csv — alternate or combined dataset (if present).

Place any additional CSVs in the project root or point your notebook to the appropriate path.

Quick Start

1. Activate environment and install dependencies (see Requirements).
2. Run the notebooks for exploration:

jupyter notebook
# or
jupyter lab

3. Run the demo app (if app.py is present and runnable):

# Example: run the script directly
python app.py

# If it's a Streamlit app (check header); to run Streamlit:
streamlit run app.py

Adjust the command depending on the app type; inspect app.py for details.

Notebooks

• rossman.ipynb / rossmann.ipynb — exploratory analysis and experiments.

Open them with Jupyter to explore preprocessing, feature engineering and model training steps.

Reproducing experiments / training

1. Ensure dependencies are installed.
2. Open the main notebook used for training (look for train.csv usage).
3. Follow notebook cells sequentially and re-run training/evaluation cells.

For script-based training (if present), run:

python train.py

(If train.py is not present, use the provided notebooks to run training.)

Notes and tips

• Keep datasets in the repository root or update notebook script paths.
• If you encounter ModuleNotFoundError, ensure the active Python interpreter matches the environment where requirements.txt was installed.
• If datasets are large, consider using a sample subset for quick iteration.

Contact / Next steps

If you want, I can:

• Run the notebooks and produce summary outputs
• Add a small requirements-dev.txt or environment YAML
• Convert a main notebook into a reproducible script

Author

Linkedln
GitHub