That is a statement not a question.
I have been at the 1Spatial user conference at Homerton College in Cambridge today. Several people commented on the difficulty that they had in finding the conference centre on Google Maps which matched my own experience. Search on Google for Homerton College and you will get directed to the centre of the Addenbrookes Hospital complex famous for heart transplants but not to my knowledge a conference venue.
The “A” marker is what the search returns while the green arrow is the location returned by Google for a search on the postcode CB2 8PH which is what the conference web site quotes and which turns out to be correct. But that is a bit confusing as there is no Homerton College at the location of the green arrow (if you look very carefully you might see a slight shading that could represent a building but definitely no attribution and certainly not the detail of Addenbrooke’s. Part of the problem is that somehow Google have the wrong postcode recorded for the college. No wonder people searching on Google were a bit confused, at least one was last heard of still driving around the hospital.
I am not in any way criticising Google, the fault, if any, is with their data supplier Tele Atlas. Homerton is not exactly a new build, it has been on the current site since 1894 so one would have thought that the same vehicles that drove round the Addenbrooke’s site could have nipped a few hundred yards up the road and recorded Homerton or at least put some carto text on the map if they were unable to capture all of the detail.
You can guess where this goes next. I looked up Homerton College on OpenStreetMap and wonder of wonders
Beautiful isn’t it? I thought so and when I showed it to one of the people complaining about Google/TA they were impressed but asked “why would anyone do that for nothing?” Well that’s another story but thanks to ito World you can see who mapped Homerton College and when here.
Now Homerton College Cambridge is not the whole of the UK let alone the world. Perhaps the demographics of a university town mean that it will have a strong team of active mappers. But it does prompt some questions about completeness, detail etc.
3 thoughts on “How good is OpenStreetMap”
Steven and I remembered when Homerton was one of the few sources of female company in Cambridge in the 60s. Any male undergraduate could have directed you!
But, on the same theme, try finding post offices. Earlier yesterday morning I was in Sheffield. Could I find the main post office? Google didn’t know; Sheffield based friend didn’t know; Big Issue seller sent me to the sorting office; local cobbler sent me to the ex post office – which was imposing; finally the third passer by was actually able to direct me to it. No decent signage and even the counter staff admitted that customers were having difficulty finding it.
When I got home I tried Google for Cambridge and Huntingdon Post Offices – why does the main post office in Cambridge come 4th or 5th down the list after suburban branches? Why does Huntingdon apparently not have a central post office at all? Is it because it has just moved into Smiths? or is it because I don’t type it with capitals?
Answers on a post card (sorry, email) to OSM please. Actually – they already have the right location for Sheffield; they have both the old and new locations for Huntingdon; but the Cambridge office still comes 4th down the list.
OSM 2.5: Google 0.5
Perhaps our new National Address Gazetteer will get all three spot on?
Nice one, it also goes to show that OSM is becoming more than just a street map.
The main issue with OSM is (still) completeness, and the main issue with the commercial street mapping is (still) the accuracy of their geocoding. With the latter I still find points of interest often marked in the (very) wrong place e.g. at the centroid of the postcode area rather than the actual location. With OSM the problem is in its inability to offer blind reliability. In an area you know relatively well OSM is quite usable because you can judge its reliability and use it with a pinch of salt. In an area you don’t know – good luck.
Time will probably fix both these issues. And time is not free, whichever way you look at it. So commercial opportunities in good quality, correctly geocoded points of interest on a complete and accurate streetmap should remain… Well, until Google and Bing derive a highly complete and consistent dataset from their own street imagery & mapping, that is. And there it seems only Navteq is left in the running, as a supplier to Bing (Google will probably do it themselves). So OSM and TA might have to come up with more creative business plans to survive and flourish beyond niche areas.
Yes Cambridge has long been a fine example OpenStreetMap mapping. Three years ago I wrote a chunk description on the ‘Places’ wiki page and Cambridge was up there. It’s also highlighted on bestofosm.org.
We have Yahoo! imagery coverage over Cambridge (which is strange actually considering there is non over Birmingham or Leeds), so that gives the city an unfair advantage, but it’s also a technology centre and more importantly and cycling centre! In fact the city was mapped fairly well in the early days, before Yahoo! declared their imagery available to us. We had a few early adopters there including a Mr David Earl.
itoworld’s “OSM mapper” tool can be good for seeing who’s contributed patches of map over a large area, but it’s only based on last user to touch an object. To see who originally mapped the Homerton college building, you’re probably better off looking more directly at the history of that element in the OSM data. If you scroll down, you’ll see the first version was initially created by non other than User:davidearl, our Cambridge super-mapper.