Predictive Weather Analysis and Forecasting System
Akshata R H 1 , Dr. Md. Irshad Hussain B 2
1Student of UBDT College of Engineering, Davanagere
2Assistant Proffesor, Department of MCA, UBDTCE, Davanagere
ABSTRACT - This project introduces a Flask-based Predictive Weather Analysis and Forecasting System that merges data from the OpenWeatherMap API with machine learning techniques and interactive visual analytics. The platform lets users "register" securely, log in, and maintain a personalized search history that provides both current conditions and five-day forecasts.
Incoming JSON responses are converted into structured Pandas DataFrames, enabling the application to compute daily averages, analyze hourly weather patterns, and generate predictive insights. By leveraging RandomForestClassifier for classification tasks and RandomForestRegressor for numerical predictions, the system can estimate "rain probability," temperature fluctuations, and humidity trends. LabelEncoder is employed to properly encode categorical features like wind direction, ensuring better accuracy.
One of the key strengths of the platform is its ability to produce "data visualizations" through Matplotlib and Seaborn. Users can explore charts that highlight temperature, humidity, wind speed, pressure, visibility, and even correlation analysis. Rendered in Base64 format, these visual outputs are directly embedded into the web interface for an enhanced and intuitive user experience.
The application also integrates authentication and data storage features, where passwords are encrypted using SHA-256 and user history is maintained in JSON files. Planned upgrades include switching to bcrypt for stronger security, migrating to a more structured database, and adopting advanced time-series models to improve forecasting accuracy.
In summary, this work demonstrates how Flask, combined with machine learning and "visualization tools," can create a scalable, reliable, and user-friendly system for weather prediction and analysis.
Keywords — Machine Learning, AI Predictions, Weather Prediction, Weather Forecasting, Search History, API, 5 Days prediction, 5 hours Prediction, Graphs.
