On 21 and 22 November, Île-de-France Mobilités organised the AI and Mobility Hackathon, with the aim of exploring the uses of AI for mobility.

For two days, the 10 teams formed " fiddled with" mobility data, manipulated it, coupled it with weather data, with event agendas, used it in machine learning algorithms, exploited the possibilities offered by language models (LLMs)... to end up with prototypes, most of which are functional, presented to the jury.

These data from Île-de-France Mobilités, supplemented by specific datasets provided by the RATP, the SNCF, and the Grand Paris Seine et Oise urban community, were for the most part unpublished. They allowed the teams to imagine new services for travellers, new tools for agents, and to take up one or more of the challenges we had set them:

       ● Improve the accessibility of mobility services;

       ● Build an AI toolbox to accelerate the development of AI for the benefit of users;

       ● Improve forecasts for mobility;

       ● Personalize the user experience of digital services to the traveler.

In addition to these challenges, there was a final cross-cutting challenge to explore the issue of the frugality of AI systems, with the aim of understanding it from the perspective of improving projects, in particular through the design of solutions that could potentially be deployed locally.

Accessibility in the spotlight

Among the most notable elements of this hackathon is the large number of teams that have taken up the challenge of accessibility and inclusion, as well as that of personalization. We will also note the diversity and quality of the participants' profiles, as well as the level of technical work carried out by the teams, whether on data, algorithms, or the use of LLMs.

The implementation of technical resources upstream...

A special effort has been made to provide the teams with data and resources to support the development of prototypes.

A datalab was deployed by Île-de-France mobilités for the hackathon, which is an instance of Onyxia, an INSEE product distributed under an open license that relies on " cloud " technologies while limiting adherence to hosting solution providers.

Through Onyxia, the teams had at their disposal python and SQL development environments , an Elasticsearch vector database and the possibility of instantiating and prototyping chatbots (with Ollama and OpenWebUI). The datasets themselves could also be put online on the Datalab (by an S3 - minio storage service). Finally, IDFM has made access to generative AI services available to candidates via team-based access keys.

To encourage the emergence of ideas and proposals, upstream work was carried out to identify data of various types: long and detailed time series (historical train transit times communicated by SNCF Transilien, validation data, aggregated route search data, etc.), textual passenger information data, real-time API of the PRIM platform, location information such as the position of lifts, stairs and call terminals (RATP data), etc.

By targeting relevant datasets by challenges, we sought to facilitate the start-up of the teams and their projection in the choice of their solution.

… and operating results...

One of the objectives of the hackathon was to identify in a pragmatic way the keys to providing operational services to users thanks to AI.

The possibilities of classification, text transformations, and forecasting, which had been identified prior to the hackathon by IDFM, were well highlighted, in particular by the projects aimed at simplifying and contextualizing passenger information.

The teams were also interested in the multimodal nature of generative AIs (ability to process a variety of sources: images, texts, sound, videos, etc.) to imagine dynamic and fluid uses for the benefit of effective information for passengers, particularly when they are cognitively disabled.

Thus, in addition to the techniques for calling LLM services and the PRIM APIs, several teams have worked on the implementation of ergonomic passenger information solutions, often using audio (whisper, local text2speech, and OpenAI API) with a concern for responsiveness (use of websocket technologies). Solutions using image analysis have also been demonstrated, in particular for the simplified and contextualized interpretation of traffic information display screens.

… which allow us to imagine the challenges of tomorrow for IDFM.

This work highlights the wealth of perspectives offered by the multimodality of language models , which offers direct operational means to allow passenger information to be adapted to the multiplicity of types of users and uses of public transport.

To activate these opportunities, one of the challenges for Île-de-France mobilités will be to assume its role as a regional platform for the centralization of quality data.

These multimodal AIs would indeed benefit from being fed by specifically annotated datasets that do not yet exist on a regional scale and that will have to be built such as images, sounds in stations, in trains or station maps.

It should also be noted that other data needs have been expressed by the candidates and would motivate a centralization work:

  • detailed data relating to the maintenance of the lifts, their environment (hygrometry, ambient temperature, etc.), and their use, to set up predictive maintenance models;
  • data on operational incidents and their causes, to develop forecasting models;
  • or data relating to the presence of agents in stations and stations, and the location of call terminals to propose use cases around the feeling of security.

In a context of rapid acceleration of information processing technologies (around AI and data), the format of a Hackathon as a collective effort to federate ideas and resources, constitutes an interesting contribution to the maturation of a collective vision. It allows public policy issues to be highlighted and allows them to be taken into account at a key moment in the adoption of technologies and their implementation on an industrial scale.

This format is also an opportunity to animate a community of experts who, in their daily work, are ambassadors of the collective vision and the necessary improvements observed collectively.

Shared software bricks

In a logic of capitalization of future AI projects of IDFM and the mobility ecosystem, all the developments carried out have their Open code and Working code snippets are highlighted to encourage inspiration and reuse.

While we can't mention all the potential reuses, we can nevertheless highlight the bricks proposed by the Accit/FALC team that make it possible to " translate " content into Easy to Read and Understand. A common to be developed for as many people as possible, beyond mobility services!

You too can share your data reuse with us by clicking on this link.

Of course, we remain available to receive your comments and answer your questions at [email protected] and on the Slack. 

Volunteer participants and a mobilized organizing team!

Once again, we would like to thank the 55 participants divided into 10 teams made up of both Île-de-France Mobilités partners (transport operators, service companies, etc.) and individual participants (AI students, mobility experts). They have been voluntary, very dynamic and fertile in production and exchange. All the teams brought their energy, their approach, their desire and their involvement in the challenges to be met. We would also like to thank all the experts who supported the teams during these two days.

Thank you also to the Hackathon jury who accepted the difficult responsibility of evaluating the projects : David Assouline (SNCF Transilien), Mounia Latrech (RATP), Jean-Daniel Alquier (GPSEO), Jules Pondard (DINUM) and Hélène Brisset, IDFM's digital director.

Finally, we would also like to thank Delphine Bürkli, Mayor of the 9th arrondissement of Paris and administrator of Île-de-France Mobilités, for her visit to this event and the many exchanges she was able to have with the teams.


Overview of the 10 proposed projects

Mobil'IA 🥇 project (First prize), led by Alexandre Le Ray, Jérémy André, Alexandre Meyer and Pierre Farret.

Mobil'IA is an inclusive application that uses AI to make public transport accessible to people with disabilities.


mAIdfm 🥈 project (Second prize), led by Simon Aharonian, Nawel Boumerdassi, Corentin Barand, Achille Popelier and Zakaria Hammal.

mAIdfm is a project focused on personalizing the use of transport applications through artificial intelligence. It aims to transform the public transport experience through:

  • Personalized recommendations based on user clustering.
  • Gamification with badges, leaderboards and fun challenges.
  • Smart notifications about inspired traffic, weather or local events, tailored to user preferences.

" Our project focused on the use of artificial intelligence to offer each Île-de-France Mobilités passenger unique, personalised and fun content, based on their profile, favourite routes, and transport habits," explains the team.

" It was an intense experience, but very rewarding ," says Achille, a member of the team, " The time pressure was there, but it motivated us to give the best of ourselves. ".


Accit/FALC 🥉 project (3rd prize), led by Gaëtan Bloch, Marc Bresson, Etienne Vautherin, Cécile Balsier, Katia Awona and Théo Lartigau.

Accit/FALC is a personalized application, accessible online and offline, that provides the necessary information at the right time, on the right channel and according to the identified traveler profile and disability. This project is trying to implement the FALC format, Easy to Read and Understand. This application is coupled with an API that centralises a text simplification and scoring service, which potentially makes it a project that can be deployed and reused beyond the scope of mobility in the Ile-de-France region.


Tranquili'score 🍏 project (Frugality Award)

The aim of this project is to determine an index of station and route tranquillity for passengers on the network. This project was imagined by Rémi Coulaud, Judicaël Leger, Vincent Bories, Antoine Greaume and Loane Cotellon.


Elevate_us is an accessibility-focused application for calculating accessible routes that take into account the operating status of elevators and escalators with a user-enriched database.


Delay Forecasts Project, by Mehdi Dahoumane, Ibrahim Sobh, Alan Adamiak, Ramdane Mouloua


It is a model for predicting potential delays according to the expected conditions (weather, events, traffic information message) making it possible, as a target, to minimize  the waiting time at the station for users. The participants were able to tackle the difficulty of analysing the data on passage and delay (how to define a delay? how to take into account the delays made up?), and the weather data (which variable to use? completeness of the data for certain stations in the Ile-de-France region).

An ambitious project that highlighted the difficulty and challenge of having sufficient data in quantity and quality to activate the potential of predictive AI technologies.


MobiWise project directed by Yohann Ciurlik, Arthur Samy, Christophe De Bast, Emilie Geoffray and Daniel Breton.

MobiWise is a toolbox to improve the passenger experience in public transport in Île-de-France during multimodal journeys.


Mob'IA project directed by Mathilda Rosset, Théophile Molcard, Calvin Paumier, Jean-Charles Fournier, Michel Kaddouh, Jocerand Ducroux

A conversational agent dedicated to finding directions that is easy to use by voice.


Alter ego project directed by Caroline Muraire, Saoussen Ben Babis, Andres Ladino, Antonio Villarral, Florian Gicquiaud, Blandine De Leiris and Gaël Garcia

Alter ego is an application based on the preferences of users, especially people with reduced mobility.


Ivoice project directed by Anand Badrinath, Eva Hoyau, François Lemeille, Caspar Longin-Dimanche, Léo Da Silva, Aurélie Chantelot and Enora Préault.

Customization of audio passenger information in the event of an incident in stations.


PRIM news in your mailbox every month

PRIM news and upcoming data available in your mailbox every month

Subscribe to the newsletter