Over the past 7 weeks, Mapbox Data team completely concentrated on aligning and improving the road network of Taiwan. With regular interaction with the community in our mapping ticket and a series of diary posts about our findings, progress and learnings, we have reached the finish line of the task.

Area Covered

Current Status

Using updated Mapbox satellite imagery and Strava heat map and different offsets with respect to Bing imagery, the data team has aligned 64,680kms of road network in Taiwan which forms the 54.4% of entire Taiwan road network. The overall project consisted of 43 tasking manager tasks covering 27,000 sq kms which is 74% of total area of the country.

Mapping Progress

Crossing highways and buildings cleanup

During our alignment work we also noticed some building tracing work from local university group which needs local knowledge and cleanup from the community side. To make the cleanup work easier, we have identified these overlapping features in Taiwan and added it as a to-fix project for the community to review and fix.

Total of 3,132 buildings are in overlap with highways - 2,862 with minor-roads and 270 with major-roads

Community Support

The Taiwan community were very responsive all the time by giving us on ground inputs which helped us throughout the project. We thank everyone, especially @Supaplextw, @assanges and @kagami3421 for your support and guidance and hope the collaboration and involvement continues in all our future projects.

Thank you,

From Mapbox Data Team.

Over the past few weeks, Mapbox Data Team has been reviewing the OpenStreetMap Data quality in Taiwan with respect to Satellite imagery. By running a comparison against Bing Imagery and available GPS traces like Strava Heat Map we were seeing a notable data offset of 15-20 meters with respect to GPS Traces.

Data offset

Some observations

• City centers have varying offset which is not aligning to GPS Traces.

• Some major highways running between cities are having huge offset.

• Roads in mountain areas are not matching to Strava traces

• Some motorway junctions don’t match with the GPS Traces

Community Involvement

• With respect to the offset and data realignment, there have been a few discussions from our side in Taiwan mailing list and Taiwan OSM group in Facebook and members of Taiwan community have expressed their support in fixing these alignment problems.

New Mapbox Satellite for Taiwan

Seeing these imagery offset issues, Mapbox Satellite is now rendered with fresh imagery for major cities in Taiwan.

• New Mapbox Satellite imagery samples

Mapbox and Strava aligning compared to Bing Imagery

• Mapbox Satellite aligning with Strava in misaligned motorway junction

These are the cities where new Mapbox Satellite imagery is rendered. With respect to data misalignment and improvement, we hope Mapbox Satellite imagery will now enhance mapping in Taiwan!

Subway systems are usually operated by a specific agency or are unified under a single network name. By mapping the railways stations in a system using a consistent network or operator tag, it is possible to provide a unique identity to those stations or lines on the map, just like a local person is used to.

The recent roll out of Mapbox Streets V7 included custom metro icons for 22 major subway networks based on these OpenStreetMap tags. This makes the map more usable by providing a familiar visual connection with their environment.

You can check out these customized icon on this style.

Adding missing operator and network tags

If your city has a major subway or transit system, you should consider checking the operator and network tags for consistency so that it can be supported for customized icons. To do this, you can follow this workflow that we use in our data team to identify stations that require these missing tags.

Inconsistencies in Barcelona

One of the major systems where we found tagging inconsistencies is Barcelona.

Overpass Query: Barcelona Metro Lines

As per Wikipedia of Barcelona Metro there are total of 180 subway stations, but from Overpass we get the following:

• station=subway with network or operator tag
  • network tag : 156
  • operator tag : 79
• station=subway without network or operator tag : 181

Note : The stats number from overpass is greater than official number as the metro entrance and the metro stations are not tagged seperately which lead to multiple occurence of railway=station and these multiple nodes which don’t count for the overall coverage. With the current mapping scheme in Barcelona it is not possible to get exact coverage of 180 station.

Update your city!

There are more cities around the world where these network and operator tag coverage is not good and need local mappers attention to tag them. Its important to keep in mind that the tagging of operator/network tag should be consistent throughout the system, please consult with your community or mailing list before making any large changes. This will allow us to support many more cities in the future.

Happy Mapping from the Mapbox data team!

Any path which is intended for dedicated foot traffic is considered a footway on OSM and tagged as highway=footway. Sidewalks that run parallel to the carriageway on one or both sides of the road and foot crossings at the junctions are sometimes mapped as seperate ways to aid in wheelchair routing.

Such ways need an additional footway=sidewalk or footway=crossing tag that can be used for specialized rendering.

Unfortunately many of sidewalks and crossings on OSM don’t have these tags and we have been trying to detect where they are as part of the sidewalk-tagging project.

Check out the OSM Sidewalker Map

In the first round we added the missing sidewalk and crossing tags for major cities in US which have a high activity of Strava users who are using maps powered by OSM.

The project included adding footway=sidewalk for highway=footway which runs parallel to the carriageways and footway=crossing for highway=footway which is used for pedestrian crossing.

The number of missing tags added by the mapbox data team:

There are more sidewalks in the rest of the world that could use your love. Take a look at our mapping documentation and follow our workflow to add the missing sidewalks tags in your city.

It would be great to hear from you how to make such tools more effecient and easy to use, do share your ideas and feedback that we can incorporate in our future mapping projects.

As part of the Japan Road improvement series, we just completed the very important Mega Nagoya where the task description was to align all the major roads according to orthorectified imagery from GSI using the tasking manager of TeachOSM.

The Nagoya task included

• Realigning of major roads (motorway, trunk, primary and secondary)
• Create Dual carriageways and highway_link roads with appropriate oneway tags

Nagoya was a well mapped urban area where the roads were very dense and the alignment was done by 23 excited mappers who finished the task in the period of 7days.

Next in the pipeline of improving roads in Japan is Osaka.

Osaka is a well populated second largest metropolitan city in Japan. The roads here are heavily mapped and the data density is huge. The objective of Osaka mapping task is

• Aligning the major highways(motorway, trunk, primary and secondary).
• Creating dual carriageways wherever needed with proper oneway tags.

Visit task page on teachosm, http://tasks.teachosm.org/project/105 to follow the instructions on how to map and join your hands in improving roads in Japan. The instructions are also available in Japanese, Spanish, Russian and English.

For more information on why road alignment is needed in Japan, you can read dairy post by PlaneMad which states different data issues in Japan.