Data Quality - What counts as a Trail?

[alongthepike]

Many jurisdictions have trails in their data files that many might question whether they are really trails, or more like glorified sidewalks. For instance along University Drive in Fairfax City, along Route 1 in Arlington near the Alexandria Border and along John McCormack Rd in DC. We need to work with WABA to determine criteria for what to count as a trail and then filter out the stuff that doesn't meet those criteria - either by getting the jurisdictions to remove them from their data files, or filtering them out ourselves as part of the fetch.js function.

[k-mahoney]

Related to this (and carrying on from #99), there seem to be some potential discrepancies between Fairfax County and Fairfax City bike route categorization methods in this data- specifically highlighted by University Drive, as mentioned above.

It's worth noting, however, that the primary example of this discrepancy is University Drive, a route currently seeing quite a bit of construction and name changing. Could be an outlier due to the data not reflecting these recent changes.

[alongthepike]

The Bike Fairfax map shows a much smaller subset of trails than are included in the Open Data Layer. I have asked the Fairfax GIS people if it would be possible to include a property in the open data layer that denotes which stretches of trail are considered sufficiently significant to be included on the Bike map. I think that would be a great first step toward weeding out a lot of the tiny, "not useful for transportation" purposes trails that are currently coming back in our dataset.

It would also probably help our buffer generation times which are suffering from the size and complexity of the Fairfax dataset.