[Imports] Import US County data

[Serge Wroclawski]

On Wed, Dec 8, 2010 at 11:17 AM, Ian Dees <ian.dees at gmail.com> wrote:
> With TIGER 2010 starting to roll out we need to think about how to offer
> some way of comparing OSM with arbitrary datasets.
> Here are my thoughts so far:
> 1) User uploads a shapefile
> 2) User selects attribute(s) in the shapefile to use for matching with OSM
> data (i.e. a road name or tiger identifier or something)
> 3) System performs a geospatial match (including the attribute(s) the user
> specified in step 2) on OSM data in the area of the shapefile
> 4) System renders a map with three layers: a) data in reference but not
> found in OSM, b) data in OSM but not found in reference, c) data in both
> Based on (4), mappers can modify the OSM data to more-closely match the
> geometry or meta-data of the reference data set. A more complicated version
> of this tool could allow the mapper to accept wholesale changes in geometry
> or attributes into the OSM dataset (removing or manipulating the existing
> data).

Ian,

Since we'd talked about this a few days ago, some of the code to do
this is already done for other datasets. In addition there is code to
do more general comparisons. My general feeling is that we should be
replacing as little as possible, and focusing on where the datasets
diverge, and then offer up something like "musical chairs" as
comparisons.

I'm also wondering (and I haven't seen the dataset to know) what the
data looks like in terms of improvements in road clasificiations, as
at least in the DC area, I found more areas of inaccurate
classifications than I did in geometry.

I think we should get together a group of interested parties and get
something working on this, hopefully re-using some of the existing
tools that have been developed to solve similar problems.

tl;dr Let's solve this problem as a small group.

- Serge
> This is all pretty tricky, but I'd love to hear other people's opinions.
> On Wed, Dec 8, 2010 at 10:06 AM, Mike N. <niceman at att.net> wrote:
>>
>> I have obtained a set of road centerlines from a US county that has been
>> published in the public domain.   I would like to use this to correct the
>> OSM roads for that county, since the TIGER data was one of the "poor
>> alignments" that we are so familiar with.   In addition, the road names have
>> all been changed for E911 consistency.   This file is likely even newer than
>> the 2010 TIGER centerlines.
>>
>>  I have used Ian's shp-to-osm to begin looking at the data.
>>
>>  1.   Does anyone have a rules file for a typical municipal road
>> centerlines shape file conversion?     I started to develop this, but can't
>> believe no one has done this before.
>>  2.   Is it good to preserve TIGER tags for untouched roadways, or just
>> delete and replace them?   The geometry is so far off that it is easier to
>> delete and replace.
>>
>>  Usual import disclaimers -
>>
>>  Any relations will be preserved.
>>  Any existing road alignment edits will be preserved (such as for
>> interstates)
>>  Any existing attribution will be preserved (speed limits, traffic lights,
>> etc)
>>  Corrected data will be manually stitched to existing data at borders or
>> pre-edited roads.
>>
>>
>>
>> _______________________________________________
>> Imports mailing list
>> Imports at openstreetmap.org
>> http://lists.openstreetmap.org/listinfo/imports
>
>
> _______________________________________________
> Imports mailing list
> Imports at openstreetmap.org
> http://lists.openstreetmap.org/listinfo/imports
>
>