Leveraging Python for data modeling, visualization and remote sensing of fire hotspots in the Amazon: a case study of Santana do Araguaia, Brazil
Keywords: Amazon region, Python, data modeling, fire risk, fire hotspots
Abstract. Conservation of the Amazon rainforest is essential because it plays an important role in global biodiversity and climate stability. In this work, python was used in data modeling, visualization and remote sensing for analysis of fire hotspots generated from the region where Santana do Araguaia is located. With the use of shapefiles and csv data and different Python libraries, like pandas, geopandas for geographical functions and libraries with graphics packages such as matplotlib, seaborn and rasterio. Methodology encompasses a series of analytical techniques to explore the relationships between environmental variables and fire risk. These techniques include generating linear regression models to study specific correlations, creating choropleth maps to visualize spatial patterns, mapping fire hotspots to identify high-risk areas and using 2d regression graphs for detailed analysis. In addition, time series analyses are conducted using static visualizations to track changes in fire risk over time, offering a comprehensive approach to understanding fire dynamics. Remote sensing techniques are used to produce normalized difference vegetation index maps and detect changes in vegetation and land cover. This research demonstrates how Python tools can reinforce fire risk analysis and support forest management blueprint, providing relevant information for conservation efforts.