Weather_audit

I built a data audit pipeline to verify weather app accuracy. It automatically collects OpenWeatherMap’s 24-hour forecasts and compares them with next-day actual temperatures. Using Pandas and a Streamlit dashboard, it analyzes errors over time and reveals the real gap between predictions and reality

🌦️ Automated Weather Accuracy Auditor

"Is my weather app lying to me?" > An automated data pipeline that tracks, logs, and visualizes the accuracy of weather forecasts in Colombo, Sri Lanka.

---

📌 Project Overview

Companies and individuals rely on external data providers (APIs) for critical decisions. But how reliable is that data?

This project is an end-to-end ETL (Extract, Transform, Load) Pipeline that:

1. Extracts real-time temperature and 24-hour forecasts daily via the OpenWeatherMap API.
2. Transforms and normalizes the data using Pandas.
3. Loads the historical log into a CSV database.
4. Visualizes the "Reality Gap" (Forecast vs. Actual) on an interactive Streamlit dashboard.

📊 Live Dashboard

I have deployed the visualization to the cloud. You can interact with the latest data here: 👉 Click to View Live App

🛠️ Tech Stack

• Language: Python 3.10
• Frontend: Streamlit (for web deployment)
• Data Manipulation: Pandas
• API Integration: OpenWeatherMap & Open-Meteo
• Automation: Cron / Task Scheduler (for daily data ingestion)

📂 Repository Structure

├── app.py               # The frontend Streamlit application
├── collect_data.py      # The backend script that fetches API data
├── weather_audit.csv    # The database (Time-series logs)
├── requirements.txt     # Dependencies for cloud deployment
└── final_chart.png      # Static export of the analysis