On July 2nd and 3rd, a Datathon was organized by the social landlord Paris Habitat. I did not know Paris Habitat or the Datathon concept, but having experienced the demand for social housing myself, I knew there was still a lot to do in this area.

TL;DR

Social housing in France is housing intended to people whose income does not exceed certain limits within the working class and the middle class. Social housing accounts for 19% of the total French fleet. It should be noted that 70% of the French population can claim, due to their income, a social housing. In Paris, the application process of social housing is particularly long. The largest social landlord in Paris decided to ask for innovative ideas to improve the life of their caretakers and tenants, from the moment they apply to their daily life inside the social housing.

Our proposal, focused on the application process of social housing, won the third place out of 17 participating groups and  also obtained the prize of the « Coup de Coeur » from Numa, jury member.

Need inspiration !

A call for fresh ideas


The main objective of the event was the creation of ideas, of concepts : they wanted to get inspired. As explained by Paris Habitat, the Datathon is named after all the data collected by them and their partners, same data that should be the very foundation of our collective thinking. Tenants, apartments, residences, neighbourhoods, and even about Paris, so many open informations available ! There were no data scientists among the participants though. We were from various backgrounds, including tenants and caretakers from Paris Habitant, with the commun point to feel that this topic was too important to pass. The good will shown by Paris Habitat was enough to motivate us trying to propose solutions.

5years

Average waiting time from application
to a social housing attribution, in Paris

17years

Unfortunate waiting time
record for a Parisian woman.

1/ 10

parisiens are
Paris Habitat's tenants

125 000

Number of Paris Habitat's
social houses in Paris

Fast forward

A two days challenge


We were given 2 days, including an introduction and conclusion keynotes. All we could use the data for, was hypothesis validation and insight crafting.

Participants were split in three themes, based on application survey : « Better Life Together », « Improve our relations with tenants » and « Improve our Energy Performance and our Ecological Footprint ».

Each group of 20-30 participants benefited from a 3 hours brainstorming / co-creation workshop, to bring out obvious problems, solutions, and build small teams.

The actual digging and proposal crafting was allowed a 24h period. Within such a short amount of time, we aimed for the essential and only this.

Picture of social housing situation in France

It is common knowledge that application for social housing are exhausting, especially for people in precarious situations. Applicants don’t have visibility on their application status. So they regularly go to Paris Habitat headquarters, or call the reception to get updates. In Paris, it takes an average of 4 years between the application and the social housing attribution. I was astonished to learn that 30% of applications have more than 5 years of seniority.

Why ?

  • Because of a mismatch between the demand and the offer.
  • Because tenants whose family situation have changed don’t tell Paris Habitat to get another social housing (smaller or bigger)
  • Because most of the time, same places and same criteria are specified in the applications
  • Because each application is dealt with by a commission of three people at least, who gather only once a month to reflect upon every applications of their district

Such a long waiting time makes application out of date !  When social housing is attributed, applicants needs have changed: a new child, an older child gone, a new work elsewhere, different incomes…  Some extreme cases have made the headlines, like this Parisian woman who received social housing 17 years after her request.

On the Paris Habitat side, they have a given housing stock, that is too small for some types of houses. They have plans for building new houses, but those are long-term projects. In the meantime, their reception face many calls from insistent individuals who wants status updates. Paris Habitat doesn’t have enough collaborator to process applications, so when a new house is available, they take the waiting list from the oldest to the newest until they find a matching candidate. No urgency criteria is taken into account to prioritise applications.

Our Team's Design Challenge


How could we use the existing Data at Paris Habitat to enable social housing applicants to understand and optimize their demand for social housing?

Proposed Solution


The proposed solution is a web platform for application for social housing. It is free, require no registration, and preserve anonymity. This platform estimates the applicant's waiting time using machine learning on Paris Habitat's data. We called it Locaboost, and we think it addresses the most urgent problems of the application for social housing.

For those who want the short version, here is the support of our 7min pitch:

Wouldn't take much to improve a lot

The lack of transparency


The heart of the problem is the lack of transparency on the parameters of the request. Applicants are alone in front of their form, with their situation, their needs, and sometimes with emergencies too. It’s important to note that a simple change of the « where » of the application can significantly reduce their waiting time. For example, the same application can takes 5 times less time to wait if it’s in the south of Paris than if it’s in the center ! The same goes for the type of apartment.

It is possible to automatically give a realtime accurate estimation. And this data is game changer for the applicants, as well as for Paris Habitat.

Using this platform, applicants can:

  • Have better visibility on their estimated waiting time
  • Modify their application to optimise themselves their waiting time
  • Follow their estimate of waiting time over the months, like a status update

Paris Habitat also benefits from this platform :

  • There is a better rotation of the housing stock with a lot more variety in applications criteria
  • A reduction in the number of files to be processed and pending, and reduction of calls
  • An improvement in the satisfaction of the applicants and therefore the image
  • A decline in the refusal rate of applications
Our Platform

Locaboost stands on recommendation


Our story mapping has highlighted a key point of our platform : the retargeting. When a family change occurs, a better social house matching can be made. Adapting the house to the tenant’s family situation allows higher rotation of the house park, and could create opportunity for a waiting applicant.

Estimating the waiting time is the first step in the improvement of social housing experience. We give the applicant the keys to act on his own application criteria to optimise his waiting time. We do that though recommendations : according to the parameters of the application, we suggest similar accommodations with a reduced waiting time. Some people will be interested in these changes to reduce their waiting time while others will prefer to maintain their criteria, knowing their waiting time in advance.

Story Mapping


Screens


Feasible tomorrow ?


Yes !

We propose to integrate this solution within existing tools. The process is unique and managed by the French government. Our tool, anonymous and free can be advertised inside the real application process, for estimation only. The data needed to build the machine learning algorithm is available already as confirmed by an interview with the Director of Information Systems at Paris Habitat. The formula must be tuned by a data scientist.

We made a planning and cost estimation : the main job to perform will be the creation of the algorithm, the setup of the production environnement for running it.

La Team

Good Bye

Thank you for reading !


Thank you Paris Habitat for organising this successful event, to the mentors who helped us in this race against the clock. Congratulations to other competitors who have successfully challenged this.

As of today, our project is still an idea. Feel free to reach out if you want to discuss.

Thank you for reading !