Data note: flourishes and jitters

[gregallensworth]

A data note for the cleanup crew: I have noted a few segments where there are hand-jitters or "flourishes" at the end of the line drawing.

They are of a very tiny scale (0.0002 degrees, a few meters) and for display purposes they are largely irrelevant. However, part of routing is calculating which direction the end of the segment is pointing, and that would be based on the last pair of points. In cases where the end is really 3 or 4 points, very close together and effectively random, this can give strange results.

My workaround consists of two parts, and is effective:

• I use a ST_SIMPLIFY(0.0001) which shaves off some of these. (units are DD, this is some 20 feet)
• The azimuth of the path is taken of the last 3 vertices and not the last 2, in addition to the simplification above.

This workaround is effective for turning directions, but is an advisory to the data-cleanup team that this was seen. Specific pline_id entries will be given shortly.

[gregallensworth]

One specific example:

687470 -- South Fletcher Avenue

Last 4 vertices are as follows. The trail is facing southwest, but the final vertex is a tiny distance north and west from the prior, giving the end a northwest-facing azimuth.

• -81.444572 30.6009139999998
• -81.445327 30.5999120000003
• -81.446022 30.599237
• -81.4460279999999 30.5992439999998

[clhenrick]

@gregallensworth I believe Maianna identified these as well. It sounded like most of, if not all of them should be cleaned up after the GIS staff finishes topology fixes. Good to know your router will account for any left over as well.

[gregallensworth]

I have recreated this issue and supporting data elsewhere: https://github.com/GreenInfo-Network/east-coast-greenway/issues/20