The project is currently in a very early stage of development. The code is not yet ready for use. The README is also not yet complete. Contributions are welcome, but please read the CONTRIBUTING.md file before submitting a pull request.
It is not working at the moment
This project aims to predict the likelihood of an observed variable being in a certain category (biome) based on the values of the observed variables.
We aim to classify the biome of a given location based on the following observed variables:
- Bioclimactic Variables
- Temperature
- Precipitation
And in future we hope to include the following observed variables:
- Elevation
- Soil Type
- Vegetation Type
- Animal Type
More info about the data can be found here.
Predicting the biome of a given location based on the observed variables is a classification problem. There are many classification methods available, but we will focus on the following:
- Koppen-Geiger Climate Classification
The Koppen-Geiger climate classification is designed specifically for climate classification. It divides the climate into 5 main categories, which are then further divided into subcategories. The main categories are as follows:
- A: Tropical
- B: Dry
- C: Temperate
- D: Continental
- E: Polar
More information on the Koppen-Geiger climate classification can be found here.
A heavily modified version of salvah22's python implementation is used.
The Koppen-Geiger classification will then be applied to the data to create a training set for the Naive Bayes classifier.
The data used in this project is from the WorldClim database. The data is available for download here.
The 5 minute resolution data is used for this project.
The downloader script will download and extract the .tif files for the following variables by default:
- Bioclimactic Variables
- Average Temperature
- Precipitation
The data is averaged over the years 1970-2000 to create a single .tif file for each variable.
This project is licensed under the MIT License - see the LICENSE file for details
Please read CONTRIBUTING.md for details on the code of conduct, and the process for submitting pull requests.