MobilityData / gbfs

Documentation for the General Bikeshare Feed Specification, a standardized data feed for shared mobility system availability. Maintained by MobilityData
https://gbfs.org
Other
794 stars 289 forks source link

Proposal for car-sharing support #338

Closed Miryad3108 closed 3 years ago

Miryad3108 commented 3 years ago

Introduction

Transport.data.gouv is the French national access point for mobility data.

We are opening this issue on our behalf and on behalf of a working group involving the “Association des Acteurs de l’Autopartage” (AAA), a French National Association for carsharing grouping 20+ operators, and Iodines.

Why are we opening this issue?

Following European Directive EU 2010/40/UE, French carsharing operators will be requested to open and publish their data.

We have noticed strong similarities between car sharing and bike sharing. We are considering the opportunity to evolve the GBFS standard in order to support both use cases and to leverage GBFS proven efficiency and popularity.

The goal of this issue is to discuss this idea with the community, before we go forward with a PR.

Proposal overview

We would like to propose an evolution of the GBFS which allows to model car-sharing data by extending the following files.

The detailed proposal to discuss is currently in this Google Document.

Is your potential solution a breaking change?

Yes.

It seems that currently in the GBFS the share type is implicitly defined by the type of station : in case of a physical station, the share type is one-way. In case of a virtual station, it is free floating.

As the round trip share type is commonly used by car-sharing companies, the share type would become explicitly defined among one-way, free-floating and round-trip.

Note on potential variant of the proposal

Following the discussions related to the creation of a car-sharing compatible version of the GBFS specification, an alternative solution emerged. The information added is the same, but an additional optional file is added to the specification, allowing the creation of generic vehicles types, refined by a notion of vehicle sub-types. We’ll let Nicolas Frasie, of Communauto, explain this alternative in more detail.

What do you think?

We also created this issue in order to have feedback from the community about this file architecture choice. Thanks in advance for your comments.

NicolasFrasie commented 3 years ago

Hi everybody,

At Communauto, we suggest an alternative : keep free_bike_status and vehicle_types as light as possible and describe precisely the wide range of vehicles and accessories into a new file whose name could be vehicle_description.

As a result, you would obtain :

I look forward to hearing from the community about these 2 options.

Thanks @Miryad3108 and French NAP team for all the fantastic job to get there.

Nicolas Frasie

teebot commented 3 years ago

Hi. I'm representing Poppy Mobility and we support these propositions as well. It would make GBFS more malleable and usable for different car sharing scenarios. We already got the feedback from public authorities that the GBFS endpoints we currently expose are lacking such data but so far we are very reluctant to expose additional data in non-standardised formats.

teebot commented 3 years ago

What do you think @mplsmitch ? Think we could open a PR with these?

thbar commented 3 years ago

Hi @teebot! I work with @Miryad3108 at https://transport.data.gouv.fr/. Glad you support the propositions!

FWIW given the positive feedback so far, we are going to open a first PR on our side in a few days with the team. We'll comment here!

josee-sabourin commented 3 years ago

Hi all! MobilityData has reviewed the proposal in depth and produced a needs assessment. In this needs assessment we don't cover the alternative approach suggested by @NicolasFrasie, however, it does not mean we are discounting the approach. Adding an endpoint to the specification is a large step that but in this case, it could also be the correct one. How does the community feel about adding a new endpoint? Should this endpoint be required only if your system is a carsharing one or can the endpoint be valuable for other modes as well? Please let us know, we are really excited about the possibilities this new endpoint, and the addition of carsharing, can bring to the specification.

josee-sabourin commented 3 years ago

Hi everyone, since the PR has been opened in #350, I'm going to close this issue and we can move the discussion to the PR. Thanks! @teebot @alibenmessaoud @NicolasFrasie @thbar @Miryad3108 @fchabouis