The French Incubators Network

I. Introduction

We learn from out failure in order to maximize success ….

It sounds good especially if you think what is most important in characterizing an entrepreneurship mindset is.

I started a couple of months ago to focus on incubators and start-ups with a picture in my mind to identify and thus grasp the network of entrepreneurs in France. Being really interested by the topic and fan of entrepreneurship, I initially intended to draw a first mapping in order to realize if we have a dense network of entrepreneurs in France or if the migration flow of French citizens moving to Silicon Valley or London is still increasing.  I wanted to form my own opinion as regards the situation in social networks and in the meantime demonstrate the added value of these networks in acquiring knowledge on a given topic.

So i decided to scan the social realm to identify French start-ups mainly in the technology business and I had the idea of looking for in places where organizations, people communicate to gain visibility: Twitter.  This network is not the most widespread in France but for my purpose, i considered the representativeness as satisfactory for this sector.

So basically I launched a crawl on this network to collect accounts with key terms in their profile (start-up, innovation, incubator and entrepreneur) and i harvested their followings on 2 levels with the same filters:  the expected outcomes should  be a network of Twitter‘s accounts interested in entrepreneurship. Obviously, everything was done with a full compliance of Twitter terms of services.

From this perspective, we should be able to determine after a cleaning phase if the French entrepreneurs mapping has a lot of members but also connections, clusters and central accounts and at the same time we should get their subjects of interest and localization if usable.

Let’s start to focus on localization.

II. Localization

Determining the localization is critical to identify incubators in France but also their local position.

So I started to work with a network composed of 5500 accounts (13% without location filled in their Twitter profile). To calculate the origin (country and geographic coordinates) when feasible, i wrote an algorithm to compute those values taking into account the unformatted data aspect. Indeed, location in Twitter profile are free text and could have capricious forms like “worldwide”, “France”, “ÜT: 41.392914,2.179219″,  “Paris”, “Paris, France” …

  • The first result is the following breakdowns by country which illustrates the initial target to focus on French incubators and highlights the connections with other countries.
  • The second outcome just gives us the opportunity to go down in depth to localize accounts close to a geographical spot. That could be useful to identify organizations or individuals working in the startups business close to given position.

If you want to check in detail the localized accounts, go to the map here at

III. The French Connection

The second aspect i want to address is the network of connections. The idea is to analyze the interactions between accounts (“who follows who” in our case) to determine central accounts in the social organization and to outline the structure type of the network (for instance polarized, clustered, bazaar, topical).

To do this, i focused on the French network (58% of the global accounts) taking into consideration 2500 accounts with a French localization and with the highest level of centrality (i applied Eigenvector centrality which assessed how well connected an individual is to the parts of the network with the greatest connectivity. Accounts with high scores have many connections, and their connections have many connections).

I let you discover the network mapping and accounts central in this network.


  • Nodes in the Mapping represent Twitter accounts, size of nodes the value of centrality metric, edges materialize the following relationship. (Big nodes in size have high centrality)
  • Careful, loading may take several seconds and if you encounter issues, reload the page.

You can consult the interactive application through the

You can enlarge the picture here

This approach give us the chance to catch a glimpse of the network density. It’s really interesting to notice we have a complete list of organizations, individuals, companies having different roles in the start-ups landscape: incubator, startups accelerator, venture, entrepreneur’s fund, mentoring, association of entrepreneurs, magazine, blogs. This unveils the momentum in France around startups development and the will to setup various structures to promote entrepreneurship. To illustrate the situation, you can have a look on some representative examples i pickup from the mapping:

To conclude this chapter on the French Connection, i just want to highlight the influence of the US (East Coast and Silicon Valley) and London in the network. French diaspora is a reality and US structures a source of inspiration. This is materialized by the density of connections across continents.


Can we squeeze something out from the Twitter profiles ? Up to you to make your own opinion.

I decided to extract the most used terms in profile description from the 5500 accounts. I applied “stop words” to filter a little bit and i also removed the terms used to collect the data (incubator, start-up, …). The result is a words cloud depicting the state of mind of people.


We have different sub-topics that outline the entrepreneur mindset:

  • Position:  “ceo”, “founder”,”management”,”manager”,”fondateur”,”director”,”president”
  • Mindset: “passionate”,”passionné”,”enthusiast”
  • Business: “capital”,”investor”,”venture”,”business”
  • Technology: “Internet”,”geek”,”development”,”mobile”,”technology”
  • Structure: “student”,”companies”,”pme”,”projets”,”services”
  • Digital: “social”,”marketing”,”communication”,”media”,”design”,”numérique”

I hope you find this article interesting. If you look for more information about French startups, I suggest you take a look at the following websites

More Information on this topic:

Solutions & softwares: