DNP-XGboostPiendamo: Satellite-Based Building Footprints for Piendamó Municipality

Estimated building layers derived from an XGBoost classification model applied to satellite imagery for the Municipality of Piendamó. The model was developed during the DataSandbox Colombia 2020 by the Data Science Unit of the National Planning Department. The dataset is projected in MagnaSirgas 3115.

Use Cases

• Classifying land cover as building or non-building using the probability and polygon columns.
• Mapping urban development extents in Piendamó municipality using the polygon and municipio columns.
• Assessing model confidence for building detection by analyzing the probability column.
• Integrating building footprint polygons into a Geographic Information System (GIS) for spatial analysis.

Strengths

• Data is derived from a machine learning model (XGBoost) specifically applied for classification.
• Includes a probability column, which suggests confidence scores for each predicted building footprint.
• Available in multiple open formats including CSV, JSON, XML, and RDF.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Row count is unknown, which may limit suitability assessment.
• The dataset's creation date is from 2020, but the last metadata update is listed as 2026-05-18; freshness should be verified.

Provenance

Source

Unidad de Científicos de Datos del Departamento Nacional de Planeación (Data Science Unit of the National Planning Department of Colombia).

Collection Method

XGBoost classification model applied to satellite imagery.

Freshness

Last updated 2026-05-18 19:03:48; freshness should be verified.

Geography

Municipality of Piendamó, Colombia.