---
title: Budget Forecasting Gradio App
emoji: 📊
colorFrom: blue
colorTo: green
sdk: gradio
sdk_version: 6.2.0
app_file: app.py
pinned: false
---


# Budget Forecasting Gradio App

A simple Gradio app that forecasts upcoming monthly budgets for a single family using a trained Linear Regression model.

## Files
- `app.py`: Entry point for Hugging Face Spaces (Gradio app launcher)
- `gradio_app.py`: App logic; loads data, model bundle, and serves the UI
- `budget_forecasting_real_data.py`: Preprocessing + forecasting helpers used by the app
- `monthly_budget_single_family_24m.csv`: 24-month single-family dataset
- `output/best_model_linear_regression.joblib`: Saved best model bundle
- `requirements.txt`: Python dependencies for the Space

## How It Works
- On start, the app loads `monthly_budget_single_family_24m.csv`, preprocesses minimal time features, and loads the saved model bundle.
- The UI lets you pick forecast horizon (1–24) and outputs a table of predicted `monthly_budget_pkr`.

## Run Locally
```bash
python app.py
```

## Deploy to Hugging Face Spaces
1. Create a new Space (Gradio, Python).
2. Upload these files in the repo root:
   - `app.py`
   - `gradio_app.py`
   - `budget_forecasting_real_data.py`
   - `monthly_budget_single_family_24m.csv`
   - `requirements.txt`
   - `output/best_model_linear_regression.joblib` (create `output/` folder in the repo)
3. The Space will install dependencies and launch automatically.

If the model file is large, consider uploading it to a separate Hugging Face Model repo and download it at runtime in `gradio_app.py`.