I am a Master’s research student at the University of the West Indies. I am just beginning my work and am thinking about getting data from the internet (Company websites etc.), particularly fuzzy data which can be operated on to produce output precise enough to allow items to be added to OSM.
I am new to the community and was wondering if that has been done/proposed before and what are the views of the community (potential pitfalls, roadblocks etc.) regarding a project like that.
All feedback welcome. Alternative idea/modifications are also welcome. I have already checked the research and ideas pages and am familiar with their contents.
Hi, Please review the steps and cautions outlined in http://wiki.openstreetmap.org/wiki/Import . Normally imports are discouraged on the basis that if someone wants to use the imported data, they can combine it with OSM data at time of rendering or usage. But in your case, you are adding value by adding structure to unstructured data.
- Unless OSM already has good addressing data in the area you are working in, Geo-locating data obtained from the internet will be difficult. Placing POIs along the roadway has less accuracy than customary OSM POI placement. There may not be geo-location information that has a license compatible with OSM.
- Pay attention to the license compatibility of data obtained. For example, it would probably not be OK to extract from a general “Business directory” type of web site, but would be OK to individually obtain information directly from company or organization web sites.
First you should start by reading the import guidelines: http://wiki.openstreetmap.org/wiki/Import/Guidelines
Specifically the section for making sure the data license is OK. Most company websites are copyright, so you shouldn’t be copying from them to add to OSM, unless you have explicit permission.
Also discussing it with the community first.
If you do have fuzzy data (with a suitable license), then it may not be best to import it directly into OSM. Instead, you could use it as sort of checklist.
See for example this UK Bike shop locator tool: http://shaunmcdonald.dev.openstreetmap.org/bike-shop-locator/ You can mark whether or not a shop is mapped in OSM, if its not mapped, you can go out and survey it and map it in the usual way.
Thank you both for your replies, they are much appreciated. I will revise what I will do and post a new topic to the community when I find a suitable project.
Well, do you think that business directories can be transferred using this method? I have a list of b2b directory which I want to import in my excel sheet and I am using Linux platform. As you know in business directories there are information regarding clients fax, telephone, email ids, etc are given so I need those to make contact.