Inflation Prediction Project

Introduction

This project aims to predict inflation rates based on historical data. It includes a comprehensive pipeline for data processing, feature engineering, model training with hyperparameter tuning, a prediction API, and an interactive dashboard for visualizing the results.

Live Demo & Deployment

This project is fully deployed and accessible online.

• Frontend Dashboard: Deployed on Streamlit Cloud.
  • Live URL: https://sri-lanka-inflation-predictio-fxqrgkwkpklyb3iwacaf8e.streamlit.app/
• Backend API: Deployed on Render.
  • Live URL: https://global-inflation-prediction-api.onrender.com/predict

Project Structure

├── api/
│   └── app.py              # Flask API for model predictions
├── dashboard/
│   └── app.py              # Streamlit dashboard for visualization
├── data/
│   ├── predictions/
│   │   ├── predictions_notebook.csv
│   │   └── predictions_script.csv
│   ├── processed/
│   │   └── inflation_data.csv
│   └── raw/
│       └── world_bank_data_2025.csv
├── docker/
│   ├── Dockerfile.api
│   └── Dockerfile.dashboard
├── models/
│   ├── model_notebook.pkl
│   └── model_script.joblib   # Trained machine learning model
├── Notebooks/
│   └── inflation.ipynb       # Jupyter Notebook for exploratory data analysis and model development
├── src/
│   ├── data_preprocessing.py
│   ├── feature_engineering.py
│   ├── predict.py
│   ├── run_training.py
│   └── train.py
├── visuals/                  # Directory for project visuals
│   ├── project_architecture.png
│   ├── actual_vs_predicted.png
│   ├── feature_importance.png
│   └── dashboard_screenshot.png
├── .gitignore
├── docker-compose.yml
├── README.md
└── requirements.txt

Key Components

• api/: Contains the web API to serve the inflation prediction model.
• dashboard/: Houses the interactive dashboard for data and prediction visualization.
• data/: Stores raw, processed, and predicted data.
• models/: Contains the serialized, trained machine learning models.
• Notebooks/: A Jupyter Notebook detailing the entire data science workflow from exploration to model evaluation.
• src/: Source code for data processing, feature engineering, model training, and prediction logic.
• docker/: Dockerfiles for building the API and dashboard services.
• docker-compose.yml: Defines and configures the multi-container Docker application.
• requirements.txt: Lists the Python dependencies for the project.

Data Science Workflow

The core of this project is a detailed data science workflow, which is fully documented in the Notebooks/inflation.ipynb notebook. The key stages are:

1. Data Preprocessing

• Data Cleaning: Handled missing values by using forward and backward fill techniques, ensuring data integrity.
• Column Selection: Selected relevant features for the model to reduce noise and improve performance.

2. Feature Engineering

• Lag Features: Created time-lagged features for inflation and GDP growth to capture temporal dependencies.
• Categorical Encoding: Applied one-hot encoding to the Country feature to make it suitable for machine learning models.

3. Model Training and Selection

A pipeline of several regression models was trained and evaluated to find the best performer.

• Time-Aware Splitting: The data was split into training, validation, and test sets based on the year to prevent data leakage and simulate a real-world forecasting scenario.
• Models Evaluated:
  • Linear Regression
  • Ridge Regression
  • Random Forest Regressor
  • Gradient Boosting Regressor
  • XGBoost Regressor
• Best Model: The Random Forest Regressor was selected as the best-performing model based on its high R² score and low RMSE on the validation set.

4. Hyperparameter Tuning

The selected Random Forest model was further optimized through hyperparameter tuning.

• Grid Search: A grid search was performed to find the optimal combination of hyperparameters, such as the number of estimators and the maximum depth of the trees.
• Final Model: The tuned Random Forest model demonstrated the best performance, which was then used for the final predictions.

5. Model Evaluation

The final model was evaluated on the test set to assess its performance on unseen data.

Getting Started

Prerequisites

• Docker
• Docker Compose

Installation & Running the Project

1. Clone the repository:

   git clone <repository-url>
   cd <repository-name>

2. Build and run the services using Docker Compose:

   docker-compose up --build

This command will build the Docker images for the API and the dashboard and run them in separate containers.

• The API will be accessible at http://localhost:5000
• The dashboard will be accessible at http://localhost:8501

Usage

API

The API provides endpoints to get inflation predictions.

• Endpoint: /predict
• Method: POST
• Data: { "features": [...] }

Example using curl:

curl -X POST http://localhost:5000/predict \
-H "Content-Type: application/json" \
-d '{"features": [ ... ]}'

Dashboard

The interactive dashboard can be used to visualize historical inflation data and see the model's predictions. Access it by navigating to http://localhost:8504 in your web browser.

Contributing

Contributions are welcome! Please feel free to submit a pull request.

1. Fork the Project
2. Create your Feature Branch (git checkout -b feature/AmazingFeature)
3. Commit your Changes (git commit -m 'Add some AmazingFeature')
4. Push to the Branch (git push origin feature/AmazingFeature)
5. Open a Pull Request

License

Distributed under the MIT License. See LICENSE for more information.