POI-based land use classification

Dataset Details

POI-based land use datasets generated and shared by the [Geospatial Science and Human Security Division in Oak Ridge National Laboratory](https://mapspace.ornl.gov/). 
This dataset classifies land use into three classes: residential, non-residential and open space. 
The dataset has a spatial resolution of 500 meters and covers all countries and regions of the world except for the US and Greenland.

Dataset Description

Curated by: Geospatial Science and Human Security Division in Oak Ridge National Laboratory
License: cc-by-4.0

Uses

Direct Use

urban planning, transportation planning, population modeling, disaster risk assessment

Dataset Structure

This dataset has four bands. The pixel values for residential, non-residential and open space bands are probabilities of the area being the land use class. The 'classification' band classifies each pixel into one of the three land use classes with the highest probability.

Source Data

Global POI data from PlanetSense Program.

Bias, Risks, and Limitations

The POI data are not collected for US and Greenland. As a result, the land use result does not cover these two regions. The training dataset used to train the land use classification model are based on OpenStreetMap land use polygons. Some regions have better training data samples than other regions. As a result, the land use classification model accuracy are not the same across the globe. In the future, we will further improve the both the POI data and training data coverage for regions that have limited coverages.

Citation

APA:

Fan, Junchuan & Thakur, Gautam (2024), Three-class Global POI-based land use map, Dataset, https://doi.org/10.17605/OSF.IO/395ZF

Fan, J., & Thakur, G. (2023). Towards POI-based large-scale land use modeling: spatial scale, semantic granularity and geographic context. International Journal of Digital Earth, 16(1), 430–445.

Thakur, G., & Fan, J. (2021). MapSpace: POI-based Multi-Scale Global Land Use Modeling. GIScience Conference 2021.

Dataset Card Contact

[email protected]