Weather API + Weather Observations Machine Learning Exercise

This project demonstrates the use of a weather API to collect historical weather data and applies machine learning to predict future temperatures. It utilizes Python for API integration, data preprocessing, and training a k-NN regression model.

Table of Contents

• Features
• Setup
• Dependencies
• Usage
• How It Works
• Sample Output
• Future Improvements

Features

• Fetch historical weather data for the past six months using the WeatherAPI.
• Preprocess and merge historical data into a structured dataset.
• Use k-NN regression to predict tomorrow’s temperature based on recent weather trends.
• Outputs model accuracy and tomorrow's predicted temperature.

Setup

1. Clone this repository:

   git clone https://github.com/DDavis-CP/Weather_prediction_model.git
   cd Weather_prediction_model

Install dependencies: bash Copy Edit pip install -r requirements.txt Set up your WeatherAPI key: Create a .env file in the project root directory. Add the following line to the file: makefile Copy Edit API_KEY=your_weather_api_key Dependencies Python 3.8+ Libraries: requests pandas scikit-learn python-dotenv Install dependencies using:

bash Copy Edit pip install -r requirements.txt Usage Run the script: bash Copy Edit python main.py The script will fetch data, preprocess it, train a k-NN model, and output: Model accuracy Tomorrow's predicted temperature. How It Works Data Collection:

The script fetches historical weather data for the past six months using the WeatherAPI. Data includes average temperature, precipitation, and cloud cover. Data Preprocessing:

Encodes categorical features like cloud cover into numeric values. Scales numerical features for better model performance. Model Training:

Splits the dataset into training and testing sets. Uses k-Nearest Neighbors (k-NN) regression to predict tomorrow's temperature based on recent weather data. Prediction:

Predicts tomorrow's temperature using the trained model. Sample Output mathematica Copy Edit Model Accuracy: 0.88 Predicted Temperature for Tomorrow: 25.08°F Future Improvements Support for multiple locations. Enhanced model with additional features like wind speed or humidity. Integration with a frontend for live weather predictions. Use of advanced models like LSTM or XGBoost for improved accuracy.