Forest Fires in Brazil

Number of forest fires reported in Brazil divided by states

Luís Gustavo Modelli - Kaggle updated 6 months ago


business, sensitive subjects, agriculture and forestry, fire and security services


Forest fires are a serious problem for the preservation of the Tropical Forests. Understanding the frequency of forest fires in a time series can help to take action to prevent them. Brazil has the largest rainforest on the planet that is the Amazon rainforest.


This dataset report of the number of forest fires in Brazil divided by states. The series comprises the period of approximately 20 years (1998 to 2017). The data were obtained from the official website of the Brazilian government.


We thank the brazilian system of forest information


With this data, it is possible to assess the evolution of fires over the years as well as the regions where they were concentrated. The legal Amazon comprises the states of Acre, Amapá, Pará, Amazonas, Rondonia, Roraima, and part of Mato Grosso, Tocantins, and Maranhão. Data (255 KB) Data Sources amazon.csv 5 columns


  • year
    • Year when Forest Fires happen
  • state
    • Brazilian State
      • It turned out during the editing of the data that there are 3 state names starting with 'Rio': Rio Grande do Sul, Rio Grande do Norte, and Rio de Janeiro...
      • As this dataset only mentions 'Rio', I'll assume that it concerns combined data of Rio Grande do Sul and Rio Grande do Norte.
  • month
    • Month when Forest Fires happen
  • number
    • Number of Forest Fires reported
  • date
    • Date when Forest Fires where reported


  • This data consists of the number of forest fires that were "reported".
    • perhaps data not retrieved from satellite images
  • The number of heat spots that were detected by satellites above Brazil is higher than the forest fires numbers.
    • A heat spot indicates the existence of a fire in an element of image resolution (pixel) that varies from 1 km x1 km up to 5 km x 4 km.
    • Moreover, if the fire is too large, it may light other pixels, meaning that several heat spots will indicate one large fire.

The watermark extension is already loaded. To reload it, use:
  %reload_ext watermark
pandas 1.0.3
numpy 1.18.4
datetime unknown
matplotlib 3.2.1
pandas_datareader 0.8.1
geopandas 0.7.0
jupyterlab 2.1.2
seaborn 0.10.0
plotly 4.7.0
bokeh 2.0.1

the dataset has only data set by year, but there is an extra month column provided...

year state
1998-01-01 Acre 730.000
Alagoas 86.000
Amapa 278.000
Amazonas 946.000
Bahia 1224.687
... ... ...
2017-01-01 Roraima 1101.000
Santa Catarina 2354.000
Sao Paulo 2540.868
Sergipe 75.000
Tocantins 1378.959

460 rows × 1 columns