[Imports] Questions about importing data for University of Vermont campus

[Josh Doe]

Andrew,
Great to hear this! I hope you get support from the administration.

I haven't done any imports yet, but am in the early planning stages to
do so using county data, so I face many of the same questions as you.

I see the buildings have a good number of attributes, though the ones
I checked are only at version 1. I don't see a problem blowing them
away, so long as you maintain the attributes that already exist,
though I assume you'll have more to add. Be courteous though and
contact previous editors (ZekeFarwell) and let them know. As for the
roads, there aren't that many involved, so I'd say blow them away as
well. Just make sure you maintain connections between sidewalks,
buildings, roads, etc. In fact I'd encourage you to micromap UVM, by
mapping all stairs, wheelchair ramps, and ensure everything is
topologically correct so you might later take advantage of various
efforts like wheelchair routing.

As for keeping data up to date, that's trickier. If you have unique
IDs or reference numbers for objects, be sure and tag it so. For an
area like a college campus, I'm guessing manually updating will be
just as easy as an automatic method, considering even that requires a
lot of manual intervention.

For tracking changes, OWL is one, another (prettier) one is located
here: http://www.itoworld.com/product/osm/

For your purposes I'd almost guarantee you'll have to render your own
tiles, though there are a number of easy ways to do so. Being able to
have your own overlays is one huge reason, plus it'll be a real hard
sell to convince administrators to be okay with a constantly updating
map that could be vandalized (temporarily of course).

That's my bit of advice, but I certainly want to hear from others.

-Josh




On Fri, Jan 28, 2011 at 3:25 PM, Andrew Guertin <andrew.guertin at uvm.edu> wrote:
> The University of Vermont web team is working on updating our online
> campus map, and right now an OpenLayers/OpenStreetMap-based solution is
> looking like the option we'll probably choose. We have a few questions
> about importing data and keeping it up to date.
>
>
> First, some background. The University keeps quite a bit of very
> detailed geospatial and other data about the campus, as you might
> expect. However, this data is spread around various departments,
> databases and non-database files, and formats. By far the largest
> problem for the web team working on our new map is collecting this data
> and getting access to it in such a way that we can keep it up to date.
> Any technological implementation issues, in whatever framework we decide
> to use, are comparatively minor.
>
> However, assuming we have the data issues worked out, we do need an
> implementation. Our most likely choice for this (mostly at my urging) is
> OpenLayers with an OpenStreetMap base layer. To do this, we need to get
> the data into the OpenStreetMap database.
>
>
> The major questions, then, are:
>
> For an initial import, what's the best way to accomplish it for various
> pieces of data?
> We have (at least) very high quality building outline, sidewalk, and
> road data. Currently in the OSM db is incomplete and somewhat low
> quality building and sidewalk data for the campus. The road data is
> pretty good, but probably not as good as we have. For an initial import,
> can we blow away the buildings and sidewalks and replace them with our
> own? Can we blow away the roads and replace them with our own? Some
> other datasets might have similar questions (parking lots, e.g.).
> Also, how would such an import be done, technically (for a set of >200
> buildings, plus other data).
> (Link to UVM in OSM: http://osm.org/go/Zd_6Cl9d--)
>
> How can we keep this data up to date?
> Would any sort of automated process be acceptable, considering that our
> data would be both authoritative and accurate? Or would we have to watch
> for when our data changed and make changes to OSM manually?
>
> How can we watch for changes other users make to the data?
> I've found
> http://wiki.openstreetmap.org/wiki/OWL_%28OpenStreetMap_Watch_List%29
> and the rss feeds you can get from there. Is that the best way?
>
> What are our options in the case someone adds valid data that we don't
> want displayed on our base map tiles?
> For example, suppose someone adds every single emergency phone
> (http://wiki.openstreetmap.org/wiki/Tag:emergency%3Dphone) on campus to
> OSM (there are a lot), but we'd prefer to have that data in an overlay
> on our map so it can be turned on and off. Would we be forced to render
> our own tiles?
>
>
> All of this is still dependent on confirmation we can release the data,
> and, in fact, on the data itself (which we're still waiting on access
> to). Having the answers to these questions ahead of time will help make
> sure the OpenStreetMap implementation is the one we use.
>
> Thanks,
> Andrew Guertin
> University of Vermont Web Team
>
> _______________________________________________
> Imports mailing list
> Imports at openstreetmap.org
> http://lists.openstreetmap.org/listinfo/imports
>