Predictive Analysis for Sales Forecasting

This project uses Machine Learning to predict future sales based on historical data. It implements a complete pipeline, including preprocessing, model training, and deployment via FastAPI.

---

Project Structure

sales-prediction-project/
│── data/                # Stores the dataset
│── models/              # Trained models
│── src/                 # Main project code
│   ├── api/             # API code (FastAPI)
│   ├── ml/              # Model training code
│   ├── utils/           # Utility functions
│── tests/               # Unit tests
│── Makefile             # Command automation
│── requirements.txt     # Project dependencies
│── README.md            # Documentation

---

Technologies Used

• Python
• Pandas & Scikit-Learn
• MLflow (Experiment Tracking)
• FastAPI (Inference API)
• SOLID (Best practices in code architecture)
• Makefile (Task automation)

---

Useful Commands

Command	Description
make init	Sets up the virtual environment and installs dependencies.
make train	Downloads the dataset, trains the model, and saves logs.
make backend	Starts the FastAPI backend for predictions.
make clean	Removes temporary files and the virtual environment.

---

Setup

1. Set up the environment

make init

This will create a virtual environment and install all dependencies.

2. Download the Dataset and Train the Model

make train

What happens?

• Trains a regression model for sales forecasting.
• Saves the trained model in the models/ directory.

3. Run the API for Inference

make backend

The API will be available at:
http://127.0.0.1:8000

You can test the API in Swagger UI:
http://127.0.0.1:8000/docs

---

Accessing MLflow

To visualize the training experiments:

mlflow ui

Access it in your browser:
http://127.0.0.1:5000

If running on a remote machine:

mlflow ui --host 0.0.0.0 --port 5000

---

API Usage Example

Making a Prediction

Parameters:

Field	Description
1.705833e+09	Transaction timestamp
0	Payment type (e.g., cash/card)
12	Number of products in the cart
4	Type of product purchased

Endpoint: POST /predict

Example request using cURL:

curl -X 'POST' \
  'http://127.0.0.1:8000/predict' \
  -H 'accept: application/json' \
  -H 'Content-Type: application/json' \
  -d '{
  "features":  [1.705833e+09, 0, 12, 4]
}'

Input JSON:

{
  "features": [1.705833e+09, 0, 12, 4]
}

Example Response:

{
  "prediction": 200.45
}