A data analysis project that cleans, preprocesses, and explores the Algerian Forest Fires dataset. The project performs data cleaning, feature type conversion, exploratory data analysis (EDA), correlation analysis, and visualization to identify patterns associated with forest fire occurrences in different Algerian regions. README.md
Algerian Forest Fires Dataset Analysis
Overview
This project focuses on preprocessing and exploratory data analysis (EDA) of the Algerian Forest Fires Dataset. The objective is to clean the dataset, analyze fire-related patterns, and understand the relationships between meteorological features and fire occurrence.
Dataset Information
The dataset contains weather and fire weather indices collected from two regions in Algeria:
Bejaia Region
Sidi Bel-Abbes Region
Features include:
Temperature
Relative Humidity (RH)
Wind Speed (Ws)
Rain
FFMC (Fine Fuel Moisture Code)
DMC (Duff Moisture Code)
DC (Drought Code)
ISI (Initial Spread Index)
BUI (Build-Up Index)
FWI (Fire Weather Index)
Fire Class (Fire / Not Fire)
Project Workflow
1. Data Cleaning
Loaded the raw dataset
Handled missing values
Removed invalid records
Assigned region labels
Renamed and cleaned column names
Converted columns to appropriate data types
2. Data Preprocessing
Converted categorical labels into numerical values
Created a cleaned dataset for analysis
Saved the processed dataset for future use
3. Exploratory Data Analysis (EDA)
Distribution analysis using histograms
Fire vs Non-Fire class distribution
Correlation analysis
Heatmap visualization
Outlier detection using boxplots
Regional fire occurrence analysis
4. Visualization
The project includes:
Histograms
Pie Charts
Correlation Heatmaps
Boxplots
Countplots for monthly fire occurrences
Technologies Used
Python
Pandas
NumPy
Matplotlib
Seaborn
Jupyter Notebook
Key Insights
Explores relationships between weather conditions and forest fire occurrences.
Identifies highly correlated fire weather indices.
Compares fire trends across Algerian regions.
Visualizes class imbalance between fire and non-fire events.