Predictive modeling of carbon intensity with Environmental impact analysis

The project focuses on analyzing and predicting carbon intensity, which measures the environmental impact of electricity generation in terms of carbon emissions in grams of CO2 per kilowatt-hour. Using data from the UK Carbon Intensity API [14], the system tracks real-time carbon intensity and detects unusual spikes or drops, i.e., anomalies. It uses models like linear regression and random forests to predict future carbon intensity and checks for accuracy, achieving an R2 score of 0.9868 using dynamic thresholding. The project also calculates an environmental impact score, showing how environment-friendly the electricity generation is, and gives recommendations to reduce emissions based on the severity of the score, those recommendations are termed as pathways. It updates the data and graphs live every minute, making it a useful tool for monitoring carbon emissions and promoting sustainable practices over time. To enable real-time exploration, the project has integrated a dashboard in the Panel framework and hvPlot for plotting. Panel is used for creating an interactive UI and hvPlot for plotting. All the live outputs and results, including Environmental Impact Score, pathways, trend analysis graph, anomaly detection, etc., will be displayed in real-time.