Open lmz opened 2 years ago
data: M:\Data\GIS layers\TM2_Networks\model_net.gdb\links
Three data sources have road "name" information: OSM, TomTom, SFCTA. In the current data (V12), only OSM "name" is kept in the Standard Network and Model Network. Names in TomTom and SFCTA data are dropped from the "links" table during Pipeline, but are processed (along with lanes data from these data sources) and stored at Lasso lookup tables, using this R script and "conflation_result.csv" (output of Pipeline step4 conflation with third-party).
Some links that are missing "name" in v12 links have names from TomTom and SFCTA, therefore bringing in these will fill some missing names. Two potential approaches:
Note about the "shorten name" method: In Standard Network, road name still keep the list format from OSM conflation - a mix of strings and lists of strings. This comes from OSM SharedStreet conflation - when two OSM links are conflated to the same SharedStreet link, both OSM names will be kept in a string. "Shorten name" was called in Standard to Model network conversion to convert it to a string field.
This issue is partly resolve by methods.determine_road_name().
There are gaps in roadway links with the name missing which is a usability issue.
For freeways, need the route direction as well based on the route signing; e.g. 101 N vs 101 S. In TM1.5, this is done here: https://github.com/BayAreaMetro/TM1_2015_Base_Network/commit/ba5089a9a0c104f06dd73c7bfde2ea499c4bcf74
@yuqiww - would you mind summarizing how many are missing by facility type? Also, this may have improved with Ranch tools