Nompilo7/Wildfire-Prediction

Predicting the burnt area caused by wildfire in Zimbabwe. # Wildfire-Prediction Predicting the burnt area caused by wildfire in Zimbabwe. #Introduction Wildfire Prediction Challenge (Zindi) Each year, thousands of fires across the African continent contribute to climate change and air pollution, impacting ecosystems, health, and livelihoods. These fires, both natural and human-induced, release significant amounts of CO2 and smoke, exacerbating global warming and degrading air quality. Understanding the dynamics that influence where and when these fires occur is crucial for predicting their effects and managing their environmental impact. #Objective Develop a machine-learning model capable of predicting the burned area in different locations across Africa from 2014 to 2016. This model will help in understanding the dynamics of fire occurrence, enabling better prediction of future fire patterns under different climatic conditions and aiding in the development of strategies to mitigate the environmental impact of fires. Architecture diagram #ETL process Data extraction: Datasets (variable information,training and testing) in CSV format were collected from the Zindi website. To facilitate efficient data handling and avoid daily upload limits, these datasets were uploaded to Google Drive and then connected to Google Colab. The Pandas library was imported into the Colab environment to enable the loading and manipulation of the CSV files. #volume: The dataset likely includes high-dimensional data across multiple years, geographic locations, and environmental factors. Transform: For both the training and test datasets, the date column was split to separate the ID and date components. From the date, the day, month, and year were successfully extracted. Subsequently, the datetime column was dropped, along with two columns (day and climate_swe) that had only one unique value, as they would not significantly impact the model's performance. Standardization was applied using the StandardScaler to normalize the features, ensuring …