An Interactive AI-Powered Tool for Strategic Dark Store Placement
🔗 Try it out now: dark-store.streamlit.app
🔗 GitHub Repository: Dark Store Feasibility Analysis
Dark Stores are closed fulfillment centers designed exclusively for online orders, enabling faster deliveries and efficient inventory management. Where should these stores be located to maximize efficiency and revenue?
This project provides a data-driven solution to analyze, predict and recommend optimal locations for Dark Stores using Machine Learning, Data Visualization and Interactive Maps.
🔹 Key Features:
✅ Predict Demand for Different Neighborhoods
✅ Recommend Top 6 Locations to Open Dark Stores
✅ Identify High-Demand Areas Needing Multiple Stores
✅ Visualize Trends with Interactive Graphs & Maps
✅ Machine Learning-Based Demand Forecasting
1️⃣ Data Collection & Cleaning
- Raw data is processed in Google Colab notebooks (included in this repo).
- The cleaned, processed dataset is used for predictions.
2️⃣ Data Analysis & Visualization
- Population growth, order volume trends and demand spikes analyzed.
- Graphs & charts provide insights into neighborhood potential.
3️⃣ Machine Learning Model
- Uses Linear Regression to predict future demand.
- Evaluated with Mean Absolute Error (MAE) & RMSE for accuracy.
4️⃣ Streamlit Web App
- Users can interact with data, view recommendations and explore maps.
| Component | Technology Used |
|---|---|
| Programming | Python 🐍 |
| Web Framework | Streamlit 🎈 |
| Data Processing | Pandas, NumPy |
| Machine Learning | Scikit-Learn 🤖 |
| Visualization | Plotly, Matplotlib 📊 |
| Mapping | Folium 🗺️ |
| Data Cleaning | Google Colab 🚀 |
git clone https://github.com/atharvbyadav/Dark-Store-Feasibility-Analysis.gitstreamlit run MainScript.pyOnly for this repo. You can change data as per your need and upload your own Data Sets for Analysis.
📦 Dark-Store-Feasibility
│-- 📂 data
│ │-- 📂 processed
│ │ │-- Merged_Pune_Dark_Store_Data.csv
│ │ │-- Pune_Climate_Delivery_Impact.csv
│ │ │-- Pune_Neighborhood_Population_Analysis.csv
│ │ │-- Pune_Online_Activity_Prediction.csv
│ │ │-- pune_dark_stores.csv
│ │
│ │-- 📂 raw
│ │ │-- Pune_Raw_Climate_Data.csv
│ │ │-- Pune_Raw_Online_Activity_Data.csv
│ │ │-- Pune_Raw_Population_Data.csv
│ │ │-- pune-ward-wise-census-2011.csv
│
│-- 📂 notebooks
│ │-- Clean_Climate.ipynb # Cleans climate data
│ │-- DataCleaner.ipynb # Processes raw data
│
│-- 📂 app
│ │-- app.py # Streamlit app
│ │-- model.py # Machine Learning model
│
│-- LICENSE
│-- MainScript.py
│-- README.md
│-- index.html # GitHub Pages support
│-- requirements.txt
- Population & order volume trends per neighborhood
- Bar charts, scatter plots & interactive graphs
- Find the best locations for opening Dark Stores
- See order volume projections
- Identifies areas where 1 store isn't enough
- Helps optimize store placement
- Forecasts future demand trends
- Improves decision-making for dark store expansion
- View existing & recommended store locations
- Zoom in for neighborhood-level analysis
📌 Algorithm Used: Linear Regression
📌 Evaluation Metrics:
- Mean Absolute Error (MAE): Measures prediction accuracy.
- Root Mean Squared Error (RMSE): Checks for large deviations.
💡 Better ML Models: Try XGBoost, Random Forest for higher accuracy.
🌍 Live Data Feeds: Integrate real-time order tracking & traffic analysis.
📊 Competitor Heatmaps: Identify areas with less competition for strategic placement.
Contributions are welcome!
Feel free to fork this repo, suggest improvements or submit a pull request.
This project is licensed under the MIT License – feel free to use, modify and distribute it.
See the LICENSE file for full details.
If you found this useful, consider giving it a star ⭐ on GitHub!