Merging New MSF Sierra Leone Village Info (Shapefiles)

Have two separate shapefiles containing new data (locations, districts, no. of households, alternate village names) for villages already in OSM and villages that have not been previously mapped in OSM. Looking for a way to get these shapefiles in to a form that would be possible to upload. There are about 700 village points with associated attribute data in the two shapefiles.

Ogr2osm seems like a viable tool for the job but am currently unsure as to how OSM would handle having a .osm file uploaded for a wide area and how it would effect data outside of the fields we are looking to update with our new data.

Any help would be great

What you are planning to do is not a mere “upload”, it is a data import. In contrast to everyday mapping activity, imports are tightly regulated in OSM. You should definitely read http://wiki.openstreetmap.org/wiki/Import and the pages linked from there before you continue; all these rules apply to any import, whether in the context of humanitarian mapping or otherwise.

Typical mistakes made with imports include importing data unsuitable for OSM, choosing the wrong tags, importing unneeded identifiers or other special information from the external source, not checking the suitability of the license, or creating double entries in OSM.

Typical workflows for shape file import often include some form of preprocessing and then uploading the data with JOSM. JOSM can also read shape files directly with the OpenData plugin.

It is good that you have already sorted your data into “inserts” and “updates”. The process of merging external data with what is already there is called “conflation” and you’ll find several approaches documented if you search for this. Under no circumstances must existing data be simply removed or overwritten to make room for an import. In many cases, manual conflation works best.

OpenStreetMap is primarily a community project, and if at all possible we’d like our data curated by local people or, failing that, at least people with first-hand knowledge of the area they are working on. I recognize that especially in a humanitarian context this might not always be achievable, however it remains a goal to aspire to. Data imported by someone from far away can never be as good as data collected in situ. Some imports therefore take the form of a “community import” where one person prepares data centrally, but others out in the field then commit that data to OSM, piecemeal and after verifying and possibly augmenting it.

Bye
Frederik

Hi there Frederik,

Sorry i should have been more clear in the initial post. The data has been collected first hand by locals on the ground as well as MSF (the part of putting this data on to osm is part of the wider process detailed in this wiki - http://wiki.openstreetmap.org/wiki/2015_Sierra_Leone_village_mapping_data)). It has been manually cleaned and should be of great value in getting a better picture of the makeup of Sierra Leones villages.

Will start looking at conflation.

Thanks very much for your help!