|
I have barely started learning about the world of OSM and JOSM and all those other programs, so sorry if this is a stupid question. I know that OSM data is updated in real-time by hundreds of users. XAPI I know does not deal with historical or deleted records. I am trying to map the spatial and temporal variation in bank branches and atms in low-income areas. We originally were going to use EPOI data for this project, but there are substantial errors and we're looking for alternatives. OSM looks good, but the point of the project is temporal and spatial variation (in Canada for years 2000 to 2011). Can I filter OSM data based on year, or is there even yearly data archived somewhere? Thanks! |
|
An interesting problem, but not one which OSM data is able to support at present. Firstly, OSM only has a very limited notion of the historical persistence of an object. Usually this is manifested by a particular set of tags associated with a particular class of objects - railways come to mind. Various other tags have been used to capture various aspects of an objects history, see for example Frankie Roberto's presentation. Secondly, most OSM data is probably far too incomplete for the type of analysis which you envisage: in some areas banking facilities will be mapped in detail, in others they might not be mapped at all. I'll use the example of pubs in Great Britain: this is a good quality data set in OSM terms (see here and here for blog posts on pubs). (Pub closures correlate well with low income areas so this is not an entirely frivolous example). There are somewhere between 50 and 60 thousand licensed premises in the country; despite our best efforts OSM has only mapped around 30 thousand (still better than 50%). In the current economic climate pubs are closing all the time, some close and re-open a few years later. It is extremely difficult to maintain the map data accurately in the face of this pace of change, and we have limited means of indicating locations which once had a pub (see for instance this blog post of mine). This is also a set of data where it would be relatively feasible to construct a history, say, over the last 40 years: there is a large amount of data in the form of directories and monthly newsletters published by CAMRA, as well as local authority data. However, we do not have the tools to do this in OSM at present. Great, thanks for such detailed responses! OSM really is an amazing database, and an amazing way to participate in the global mapping interest, but yes I've quickly realized it's limited utility for research and geospatial analysis.
(24 Feb '13, 19:11)
vputinski
1
If you find the data you can perhaps upload it to: http://wiki.openstreetmap.org/wiki/Historical_OSM (when it gets going)
(24 Feb '13, 21:47)
emj
|
|
SK53's answer is really good and I'd just add two sentences which I think sums up why OSM isn't suitable for this: If an ATM wasn't present in OSM in 2011, that doesn't mean the ATM didn't exist. It just means no-one had mapped it yet. |
OSM was started in 2004 and from my memory data was a bit sparse in the early years and as really ramped up in the last couple. So it may not cover all the years you wish to look at. history..http://en.wikipedia.org/wiki/OpenStreetMap