SupreetGha/algerian-forest-fires

Algerian Forest Fire Analysis & Prediction Exploratory Data Analysis & Machine Learning Model A data-driven approach to understanding and predicting forest fires in Algeria. ## 📌 Project Overview In this project, we delve into the dynamics of forest fires in Algeria, using a data-driven approach to both understand and predict these fires. The aim is to utilize exploratory data analysis (EDA) and machine learning to extract valuable insights from historical data and to develop predictive models that can forecast forest fire occurrences based on given meteorological features. The dataset used in this project is sourced from the UCI Machine Learning Repository, providing a rich set of variables including temperature, humidity, wind speed (Ws), rain, and other environmental factors recorded over specific periods in 2012 across two regions: Bejaia and Sidi-Bel Abbes. ## 🚀 Technologies Used - Python 🐍 - Pandas, NumPy 📊 - Matplotlib, Seaborn 📈 - Scikit-learn 🤖 - Jupyter Notebook 📓 ## 📊 Data Analysis & Modeling **Algerian_Forest_Fires_EDA.ipynb**. - Performed feature engineering to create a binary `Classes` variable (0 for "not fire", 1 for "fire") to aid in classification prediction. - Visualized data patterns using histograms, boxplots, bar charts, and count plots to understand the relationship between meteorological conditions and fire events. - Carried out correlation analysis to identify features with the strongest association with fire outbreaks, visualized through a heatmap. - Stored and retrieved the dataset using MongoDB for efficient data management. - Conducted region-specific analysis by splitting the dataset into Bejaia and Sidi-Bel Abbes regions to uncover regional fire patterns. - Created monthly fire occurrence visualizations to identify seasonal trends in fire activity. count mean std min 25% 50% 75% max Temperature 243.0 32.152263 3.628039 22.0 30.00 32.0 35.00 42.0 RH 243.0 62.041152 14.828160 21.0 52.50 63.0 73.50 90.0 Ws 243.0 15.4 …