[Talk-us] TIGER expansion bot

[Serge Wroclawski]

On Tue, Nov 27, 2012 at 2:45 PM, Alex Barth <alex at mapbox.com> wrote:
> Serge -
>
> Awesome effort, thanks for starting this.
>
> - You are planning to run this script on all TIGER data in OSM, i. e. the entire US, correct?

Yes, I have a planned bounding box that includes the entire
continental US, and then Alaska and Hawaii.

The reason to run this on the entire US is that in testing, I found
~20k ways expanded on the West Coast.

> - Are you envisioning a (geographically) limited first run with and a review phase? If so, how will this look like?

I've run this on test data from Geofabrik of the entire US. I can
re-run that at any time.

> - Will any highway data that has been modified post-import be touched? I'm trying to ask the same question as Brian May did with the '"Main St SW" to "SW Main Street"' example.

That's not quite the same question as Brian's question- and we can
discuss that question separately, but the answer is that in my
experience with the OSM data, there are lots (tens of thousands) of
ways which have been touched by one or another editor, but which still
have original names, so yes, it makes sense to expand all ways.

As for Brian's question, the problem with it is that it's a multi-part
question, so let me break it down a bit.

As you said, the original name in this theoretical example would be
"Main St SW" and is now "SW Main Street". The reason I asked for an
example of this is for two reasons:

1. I've gone through the results of the expansion bot- in detail for
some places (Maryland, Virginia and New York) and browsing for other
places, and I haven't seen that kind of thing.

2. Given that exact situation, a human editor like myself would be
inclined to expand the revised name to "Southwest Main Street"- so
this opens up a can of worms of "What is the right answer?" - in this
case, I think if the script expanded the direction tag, then it
doesn't matter if it's moved the location in the string.

On the other hand, if the new name were "Main Street NW" - then the
script would do nothing, because no direction tag of NW would be
present.

This is a tricky situation, and deserves some real consideration, and
once we come up with a consensus of the right answer, it can be
codified.

- Serge