CLIMaiTE

🌍 Clim_AI_te: Predict. Adapt. Sustain. ⚡

🔥 Inspiration

Climate change has caused extreme weather events, disrupting agriculture, transportation, and daily life. We wanted to build an AI-powered climate prediction tool that helps people understand and prepare for upcoming weather conditions. By leveraging machine learning, we provide accurate forecasts based on historical weather data.

🧠 What We Learned

Throughout this project, we explored:

Feature engineering to extract meaningful insights from weather data.
Support Vector Machines (SVM) and how it performs in climate classification.
Streamlit for building interactive web applications.
Overcoming challenges in weather data inconsistencies and missing values.

🛠️ How We Built It

We followed a structured approach:

1️⃣ Data Collection & Preprocessing

We used a weather dataset containing:
- date – Timestamp of the weather record.
- precipitation – Amount of rainfall in mm.
- temp_max – Maximum daily temperature.
- temp_min – Minimum daily temperature.
- wind – Wind speed in km/h.
- weather – Weather condition (e.g., sunny, rainy, foggy).
Handled missing values and normalized numerical features.
Converted weather conditions into numerical labels for model training.

2️⃣ Model Training & Optimization

Implemented SVM for weather classification based on input features.
Tuned hyperparameters to optimize prediction accuracy.
Evaluated the model using precision, recall, and accuracy metrics.

3️⃣ Building the Streamlit App

Designed an interactive UI for easy user input.
Allowed users to enter weather conditions (temperature, wind, precipitation, etc.).
Integrated real-time prediction visualization. ### 🌍 Streamlit App Features

Home page:

🎨 Enhanced UI & Design

Improved homepage design with a better-aligned image for an aesthetic and engaging look.
Increased font size and optimized layout spacing for better readability and user experience.

🎭 Fun Facts & Jokes

The Home Page now features a “Tell me a joke” button, displaying a random weather-related joke for a fun user experience.

🔮 Climate Prediction

Users can input weather parameters (temperature, wind speed, precipitation) to get real-time weather predictions using our trained ML model.
The app uses Support Vector Machines (SVM) to classify weather into categories like Sun, Snow, Rain, Drizzle, and Fog.

👕 Clothing Recommendations

Based on the predicted weather, the app suggests two outfit ideas for men and women to help users dress appropriately.
Example: For Rainy Weather, the recommendations include:
- Men: Raincoat with waterproof pants, boots, and an umbrella
- Women: Trench coat with waterproof leggings, boots, and an umbrella

🍲 Food Recommendations

Provides detailed meal suggestions based on weather conditions, ensuring users eat according to the climate.
Example: On a cold snowy day, the app suggests hot soups, stews, and warm drinks like spiced tea. ## 🛠️ Built With
Python – Core programming language for data processing and machine learning.
Pandas & NumPy – For handling and preprocessing weather data.
Matplotlib & Seaborn – For visualizing trends in climate data.
Scikit-learn – For implementing and training the SVM model.
Streamlit – To build an interactive web app for real-time predictions.
pickle – To save and load the trained SVM model efficiently.
Google Colab – For model training and experimentation.
Git & GitHub – For version control and collaboration.

🚧 Challenges We Faced

Time-series Complexity – Weather data is sequential, making it challenging for traditional ML models.
Data Preprocessing – Handling missing values and encoding categorical weather types.
Balancing Performance & User Experience – Ensuring real-time predictions without lag.
Model Selection – Testing different ML models before finalizing SVM.