Machine Learning in Big Spatial Data
The project was running until June 2019 and was continued as HeiGIT gGmbH. Recent information can be found on the HeiGIT website
Analyses and Utilization of Volunteered Geographic Information from the Crowd
with Spatial Analytics and Geocomputation
Vast amounts of unstructured and spatially attributed data are continuously generated and available on the web. Prominent examples include Volunteered Geographic Information like OpenStreetMap or from Social Media. With our background in GIScience we are able to extract precious knowledge from such datasets, e.g. by finding latent patterns and regularities to help you optimise your business processes. We develop quality measures as well as algorithms and methods for improving and enriching such spatial data sources for your specific needs using intelligent methods like data mining and machine learning.
Services
- Applied research in Spatial Analyses, Spatial Simulation & Modeling, Spatial Data Mining and Machine Learning
- Development and maintenance of services for Analyses and Visualisation of Big Spatial Data from the Crowd
- Business-oriented solutions for Geocomputation & Geoprocessing, Data Integration & Fusion and Sensor Analyses for heterogeneous spatial data sources
Selected References GIScience Heidelberg
- OSM Landuse/Landcover
- OSMatrix
- DeepVGI - Deep Learning Volunteered Geographic Information
- Fusing spatial data from Social Networks
- LandSense: A Citizen Observatory and Innovation Marketplace for Land Use and Land Cover Monitoring
- iOSManalyser
- Detecting Urban Areas in OSM using Machine Learning
- HistOSM.org - Historic Information in OSM
- WeGovNow: Collective and participative approaches for addressing local policy challenges (EU Horizon 2020)
- PsychoGeography
- Data quality assurance in citizen science
- Generating Knowledge for the City
- OSM Agent: Predicting OSM data
Further Projects - Publications - PhD Thesis
Further Information
