OSM Quality Mapping: Metrics to monitor Buildings outbounds
2019-09-22 , Hörsaal Ost

Mapathons and Imports represent a Quality Challenge for the OSM community. This presentation focuses on Buildings. It presents Metrics and show progress of tools development to monitor and correct Quality problems in the OSM database or before importing.


Mapathons and Building import projects represent major OSM Quality problems to monitor. Potentiel 3.0 and OSM-RDC started a project in 2018 to monitor Data quality problems. We present the work done with metrics to monitor quality problems among various Projects (see OpenDataLabRDC Blog Articles about Building Quality ). Individual building geometries are evaluated and classified (ie. orthogonal, quasi-orthogonal or irregular). Topological errors are detected (ie. incomplete polygon, self-crossing, overlaps). Geojson visualisations and Statistics let evaluate a project. Output of List of osm_id's for various problems (ie. topological error, irregular building) let's import simply into JOSM with Overpass queries.

Current development is to go a step further and correct buildings quasi-orthogonal angles in an OSM file taking into account neighbor buildings to avoid distorsions with blocks of buildings aligned on a street (See OQ_Analysis).

Programming is based on the Osmosis schema, Python and PostgreSQL-PostGIS. OSM files to analyse can be imported from various sources (ie. Geofabrik, Overpass, osm export). Each OSM file (specific date-Zone) is converted to PosGIS and stored in a specific PostgreSQL schema via Osmosis. The topological analysis and geometry measures are made with PostGIS functions.


Subtitle

How can we better monitor and correct Quality problems

Talk keywords

quality, building, mapathons, imports, postgis

Pierre started humanitarain OSM mapping with the Port-au-Prince earthquake in 2010. He is involved both for humanitarian actions, local community support and locally in Canada. He volunteered for mapping, to develop tools, to coordinate remote actions and advocate for major OSM Responses from 2012 (see Slideshare presentations). Pierre and others innovated using maps, Mobile units and diligent actions to support the logistic of the responses and develop an efficient ecosystem with the responders.
They succeeded to convince the UN Agencies and humanitarian NGO's to use OpenStreetMap as the De Facto map for emergencies.

He had the opportunity to meet OSM communities in various countries including Haiti and Sénégal. He participated in various training actions for local communities and developped OSM statistics & Map tools (OpenSource). He is the orginal author of the Highway Tag Africa wiki page. The OSM Image week1 2013 shows Minova, DRC mapped in three days while 100,000 displaced people seek refuge in the town.

More recently, while contributing with Potentiel 3.0 and OSM-DRC to develop the DR Congo OSM community and the OpenCities Project in Kinshasa, he started to develop the Building Mapping Quality Project and published various articles on the subject. The team also had to focus rapidly on the 2018 DR Congo Ebola Response. This tool helped to monitor quality problems in Butembo and react quickly. At the same time, he continues to develop the quality tool to integrate buildings orthogonal corrections.