🔗 Repository: https://github.com/anasrobo/Uber_Trip_Analysis
An interactive Streamlit app that analyzes NYC Uber pickups (Jan–Jun 2015) and forecasts hourly demand using an ensemble of three ML regressors.
- 🔍 Time-Series Feature Engineering: hour, day-of-week, month, weekend flag
- 📈 Rolling Stats: 24-hour rolling mean & standard deviation
- ⏳ Lag Features: previous 1–24 hour trip counts
- 🤖 Ensemble Learning: XGBoost, Random Forest, Gradient Boosting combined via weighted average
- 🌐 Streamlit Dashboard: Interactive UI with Plotly charts
- 🔮 24-Hour Forecast: Recursive forecasting for the next day
Uber_Trip_Analysis/
├── streamlit_app/
│ ├── app/ # 📊 Streamlit application source code
│ ├── train_and_save_models/ # 🧠 Training scripts & model saving
├── assets/
│ └── newplot.png # 📈 Sample visualization
├── notebooks/
│ └── Uber_Trip_Analysis # 📓 Jupyter notebook (EDA, modeling)
├── models/
│ ├── xgb_model.pkl # 🌲 XGBoost saved model
│ ├── rf_model.pkl # 🌳 Random Forest saved model
│ ├── gbr_model.pkl # 📉 Gradient Boosting model
│ └── ensemble_weights.pkl # 🧩 Custom ensemble weights
├── uber-raw-data-janjune-15.csv # 📁 Raw Uber data (~512 MB)
├── requirements.txt # 📦 Required Python packages
└── README.md # 📘 You're here
Uber NYC pickups (Jan–Jun 2015), timestamped at the ride level. Resampled to hourly counts for forecasting.
📥 Download (512 MB)
Google Drive Link
After downloading, place
uber-raw-data-janjune-15.csvinto the project root.
#bash git clone https://github.com/anasrobo/Uber_Trip_Analysis.git cd Uber_Trip_Analysis
python -m venv venv
.\venv\Scripts\activate
source venv/bin/activate
🚀 Usage streamlit run app.py Opens at http://localhost:8501
Actual vs Predicted trip curves (XGB, RF, GBR, Ensemble)
24-Hour Forecast dashed line
Performance Metrics: MAPE, RMSE, R²
Load & Preprocess
Read CSV, parse Pickup_date as datetime
Resample to hourly counts, set index
Feature Engineering
Time features: hour, dayofweek, month, is_weekend
Rolling stats & lag features
Model Training
Train XGB, RF, GBR on training split
Ensemble
Weighted avg of model predictions (weights learned via CV)
Evaluation
MAPE, RMSE, R² on test split
Forecasting
Recursive 24-hour ahead using latest data & features
-
Actual vs Predictions & 24-Hour Forecast
-
Input Form UI
Model MAPE RMSE R² XGBoost 12.3% 450.2 0.842 RandomForest 13.1% 478.9 0.817 GradBoost 12.8% 462.8 0.831 Ensemble 11.5% 432.1 0.858
(Actual values may vary.)
streamlit pandas numpy plotly scikit-learn xgboost joblib
Fork & clone
Create a branch (git checkout -b feature/your-feature)
Commit & push
Licensed under MIT. See LICENSE for details.
“The best way to predict the future is to create it.” – Peter Drucker