⏳ Time Series Forecasting App

A user-friendly Streamlit web application that allows you to forecast time series data using ARIMA, Prophet, and Random Forest Regression. The app visualizes and compares the forecasts with actual data, providing key performance metrics like RMSE and R².

📌 Features

• Upload your own .csv time series dataset
• Automatic preprocessing and cleaning
• Interactive parameter tuning (ARIMA order, Random Forest estimators)
• Visual forecast comparison for:
  • ARIMA
  • Facebook Prophet
  • Random Forest Regressor
• Evaluation Metrics: RMSE and R² Score

🧠 Models Used

🔸 ARIMA

A classic statistical model for univariate time series forecasting, best for stable and linear patterns.

🔸 Prophet

A robust model developed by Facebook for trend-seasonality forecasting, especially suitable for business data.

🔸 Random Forest Regressor

A machine learning model that uses lag features and statistical features (moving average, standard deviation) for regression-based forecasting.

📁 Example Dataset Format

Your dataset must contain at least:

date	target
2023-01-01	112.0
2023-01-02	114.5
...	...

• date must be convertible to datetime.
• There must be a date coulman (its not cass-sensetive because there is a method to transfer it to lower-case).
• The target variable should be numerical or convertible to numeric.

📝 Notebook

The full modeling process is documented in Model.ipynb for a (TESLA stock dataset ), which includes:

• Data exploration

• Cleaning and transformations

• ARIMA parameter selection

• Prophet tuning and visualizations

• Feature engineering for Random Forest

• Performance comparison

🚀 Deployment

This app can be deployed via:

Streamlit Cloud > https://forcast-time-series-models-app-fbrmpjhrpxepphcwbh3m8j.streamlit.app/

⚙️ How to Run Locally

# Clone the repository
git clone https://github.com/YamenRM/Forcast-Time-Series-Models-App.git
cd Forcast-Time-Series-Models-App

# Install dependencies
pip install -r requirements.txt

# Run the app
streamlit run app.py

📌 Future Improvements

• Add LSTM and XGBoost forecasting models

• Allow automatic ARIMA and Prophet parameter tuning

• Enable multivariate time series support

• Interactive plots using Plotly

👨‍💻 Author

YamenRM

💡 AI/ML Engineering Student | UP

📍 PALESTINE | Gaza Strip

💪 Stay strong!

⭐ Star this repo if you find it helpful!

📬 Feel free to open an issue or pull request.