FloodML is a web application that leverages Machine Learning to predict floods based on weather and historical data.

• Clone the project
• Install dependencies
• Activate virtual environment
• Run pip install -r requirements.txt
• Run python app.py

Floods, exacerbated by climate change, pose a growing threat worldwide. To address this, we created FloodML—an interactive web app for predicting and visualizing floods.

• Flood Prediction Plot (Red dots for predicted flood locations)
• Precipitation Plot (Bubbles indicate precipitation volume)
• Damage Analysis Plot (Bubble size represents estimated monetary damage)

• Damage Analysis Heatmap (Colors indicate predicted monetary damage)
• Precipitation Heatmap (Dark red areas signify higher precipitation)
• Flood Prediction Heatmap (Darker red spots indicate likely flood locations)

• Displays precipitation volume over Indian cities
• Uses NASA's Global Precipitation Measurement Project data

• Real-time weather forecast and flood prediction for any city
• Includes temperature, humidity, cloud cover, wind speed, and precipitation

• Scraped floodlist.com using Beautiful Soup 4
• Utilized Visual Crossing weather API for historic weather data
• Applied data augmentation techniques for model diversity

• Built on the sci-kit learn library
• Explored various models; Random Forest Classifier achieved 98.71% accuracy
• Saved model using pickle

• Integrated major Indian cities' data with weather factors
• Utilized Plotly chart studio for diverse map visualizations

• Developed with Flask framework
• Hosted on Heroku

• Limited data availability for floods in India
• Plotly integration complexities
• Git merge conflicts due to encoding and version disparities

• Created a robust dataset for accurate flood predictions
• Implemented a machine learning model with over 98% accuracy
• Successfully integrated data augmentation and visualization techniques

• Enhanced skills in web scraping, data mining, and Plotly
• Expanded proficiency in machine learning models

• Expand coverage to cities worldwide
• Implement image classification for flood detection using satellite data

FloodML aims to aid people and governments in flood preparation, potentially saving lives and livelihoods. Visit FloodML to explore the tool.