sottohy/Weather-Prediction

Regression model for tomorrow's weather in Alexandria, Egypt based on historical weather data # Weather prediction model using Machine Learning This project predicts the maximum temperature (TMAX) for the next day using historical weather data. The model is built using ridge regression and evaluated using backtesting to account for the time series nature of the data. ## Project Overview This project utilizes historical weather data to predict the maximum temperature of the following day. The dataset is cleaned, preprocessed, and a ridge regression model is trained. The model is evaluated using backtesting to simulate how it would perform in a real-world scenario. ## Data Description The data was obtained from the NOAA website ncdc.noaa.gov and the csv file is uploaded here. The uploaded data is obtained from multiple stations accross Egypt from 1993 to 2019. It contains daily weather observations, including features such as temperature, precipitation, and snow depth. The dataset is indexed by date and includes the following key columns: - TMAX: Maximum temperature - TMIN: Minimum temperature - NAME: Station name - STATION: Station identifier ## Dependencies The project requires the following libraries: - pandas - scikit-learn - pickle ## Data Preprocessing - Loading Data: The weather data is loaded from a CSV file and the date column is set as the index. - Handling Missing Values: Columns with a high percentage of missing values are dropped. Remaining missing values are handled by dropping rows with missing values. - Filtering Data: The dataset is filtered to include only data from a specific weather station. - Index Conversion: The date index is converted to a datetime type. - Target Column Creation: A target column is created by shifting the TMAX values by one day to represent the next day's maximum temperature ## Modeling ### Backtesting Backtesting is used to evaluate the model on time series data: - Training and Testing: The model is trained on a rolling basis, using 10 years of historical data and making predictions fo …