Geospatial Intelligence and Data Science

GECOsistema's Geospatial Intelligence and Data Science service line combines GIS, satellite imagery, data-driven methods and environmental models into tailored web applications that support decisions in public administration, private industry, agriculture, renewable energy and water resource management. The Rimini-based environmental engineering company sources, processes and analyses Earth observation data and applies machine-learning techniques to it, using libraries such as TensorFlow and scikit-learn on Microsoft Azure and Amazon AWS cloud infrastructure.

Stated applications include:

• Climate-smart agriculture and irrigation: algorithms fed with data from low-cost sensors, drones and satellite observations predict crop irrigation and fertilisation needs and support crop-yield forecasting; the agricultural work draws on Copernicus Sentinel data and Copernicus C3S climate services.
• Hydropower forecasting: machine-learning models predict the variables that drive hydropower production, to support energy trading, through a web interface.
• Flood and natural-hazard mapping: Sentinel-2 imagery is processed with data-driven thresholding, together with DEM-derived, land-use, lithology and other spatial indicators, to map flooded areas quickly.
• Optimal siting: GIS-based spatial decision support systems couple multicriteria analysis with map algebra to select sites for energy plants, agriculture and linear infrastructure.

For web delivery the company has developed GECOMap, a framework for building web-mapping applications based on Python, GDAL and OpenLayers that follows OGC standards such as WMS and WFS. Its most common technology stack also includes MapServer or GeoServer and PostgreSQL/PostGIS, and ESRI solutions are supported; the company also builds desktop GIS projects in QGIS and ArcMap. Each service is tailored to the client through a co-design and agile development process.