Dashboard Demo

🚀 Project Overview

This project demonstrates an end-to-end flight delay prediction system. By integrating Tableau with Python (TabPy), the system fetches live weather data from the NOAA API and runs a Logistic Regression model (trained on 4M+ records) to predict delays in real-time.

🔑 Key Features

  • Hybrid Data Pipeline: Merges historical rolling statistics with live API weather data.
  • Real-time Interaction: Users select flight routes, and the model computes delay probabilities instantly.
  • Explainable AI: Provides top-3 SHAP root cause explanations for every prediction.

📊 Visuals & Poster

Since Tableau Public does not support external Python scripts (TabPy), the interactive dashboard above utilizes a static dataset snapshot for demonstration.

View Source Code on GitHub Download Research Poster (PDF)

🛠 Technical Architecture

  • Data Source: Bureau of Transportation Statistics (BTS) & NOAA Weather API.
  • Model: Logistic Regression (Scikit-learn) and SHAP, optimized for inference speed.
  • Integration: TabPy server acting as the bridge between Tableau frontend and Python backend.