Using public data sources as initialization

[TTalex]

... Seems more complicated that expected.

The transport data commons initiative (TDCI) has developed tools and data standards to help gather transport data from different providers.

• Github: https://github.com/transport-data/tool
• Docs: https://docs.transport-data.org/en/latest/

The tool uses python and the SDMX format. It appears to be compatible with a few sources, either by fetching existing SDMX feeds (edb, eurostat), or by converting specific data. All in all, the repo has a nice listing for open transport data.

Sadly, the tools seem to be a work in progress, without updates in the last years. Installation also seems broken. So I'm unsure how usable the result would be for this tool, especially since programming languages aren't even the same...

Moreover, while the idea of converting various sources to a standardized format (SDMX), more work would be needed to identify the proper sources, and convert it case by case to be usable here. The process is far from being easily automatable, with differences in mode names, geolocations, available years ...

With that in mind, maybe a first step would be to focus on a specific provider, for example Eurostat seems to have many of the required data (for Europe only, sadly)

• Full list
• Road motor vehicle traffic performance by traffic and registration location and type of vehicle
• Stock of vehicles by category and NUTS 2 regions
• Mopeds and motorcycles by type of motor energy
• Passenger cars, by type of motor energy

Since the TDCI tools doesn't really add anything to an already compatible SDMX feed, we can try to skip it, and use a js compatible library instead, such as https://github.com/sosna/sdmx-rest4js.

However, it is unclear how to query these specific feeds using the library, especially since the Eurostat SDMX api used in the lib appears to be broken.

---

All in all, I belive there's a good idea to help users with suggested public data, but it seems to be such a case-by-case endevor, that it might be more productive to just suggest data sources instead of struggling to load uncertain information.

Maybe an easier approach would be to manually enter data for a few countries, and implement something to share values between projets (a useful feature anyway)