Written by André Vermeij, Kenedict Innovation Analytics
The combination of knowledge and pooling of resources from educational institutions, governmental bodies, research institutes and businesses is often expected to increase the pace of innovation and generate beneficial outcomes for society as a whole. As such, governments actively aim to stimulate this type of collaboration through incentive programs and subsidies on the national level.

The network of universities, research institutes, governmental bodies and businesses in the Netherlands which participated in 3 or more projects as part of the Topsectoren policy
Stimulating the triple helix: the Dutch ‘Top sectors’ policy
In the Netherlands, the ‘Topsectoren’ (top sectors) policy focuses on a total of 9 sectors in which the country excels: for example, Energy, Water & Maritime, High-Tech Systems & Materials and Agriculture & Food. The government actively invests in stimulating collaborations in this triple helix of industry, science and governmental bodies within these sectors. Ultimately, the goal of the policy is to solidify the competitive position of the country in the world economy and ensure sustainable economic growth.
In news articles and interviews with governmental spokespersons, public-private partnerships are often mentioned as key to the functioning of the policy as a whole. Seldomly, however, is this quantified and visualised to allow deeper insights into the actual collaborations which are taking place. Let’s dive into some data to provide some steps in this direction.
Identifying public-private partnerships
The data used in this post comes from a database hosted by the Netherlands Enterprise Agency (RVO). Part of the Dutch Ministry of Economic Affairs, RVO is responsible for all innovation-related subsidies granted to organisations in the Netherlands. The database contains information on the projects (e.g. titles, topics, budgets) and the receiving organisations, and thereby serves as an excellent resource for mapping partnerships. Of the 29,906 projects in the database, 5,247 are described as relating to the Topsectoren policy. These thus serve as the basis for the analysis.
Since the dataset does not include information on the types of organisations (for example, government, educational, business or non-profit), the first step was to extract all projects which have at least one public participant. This was achieved by focusing on keywords relating to educational institutions (universities/universities of applied science), provinces, municipalities, ministries, major applied research institutes (the so-called TO2 federation – TNO, ECN, Marin, Deltares, NLR and DLO) and various large other public bodies (e.g. NWO, KNAW, CBS, PBL, Rijkswaterstaat, RIVM). Finally, this leads to 1,404 projects in which at least one governmental body, research institution or educational institution participated*.
A common issue with large-scale data sources is that organization names are often written in various ways. This is no different here – a combination of fuzzy matching and manual cleaning was thus applied to make sure the analysis excludes as many duplicate entries as possible. After cleansing, this leads to a total number of 2,355 identified unique project partners.
Mapping the Triple Helix
As a start, let’s look at the full network across all top sectors, projects and participants. To not clutter the visual too much, we’re focusing on participants who appear 3 or more times in the dataset, plus their related projects. The larger a participant, the more project it has participated in; colours signify clusters or communities of participants and projects and are assigned by a community detection algorithm:
You can click any project or participant to view further details in the panels on the right, and search for keywords of interest using the search box on the left. A first glance at the network shows that research institutes TNO, ECN and DLO, as well as the universities in Delft, Eindhoven and Twente are most connected in terms of the number of projects they have participated in.
The overall activity is clustered pretty clearly: the Agriculture & Food sector is mostly represented in the green cluster at the right, with Wageningen University and DLO playing key roles. The left-hand side of the network contains significant activity in the High-Tech Systems & Materials sector, with various shared project participations between technical universities and corporates active in high-tech such as NXP and ASML. The activity at the far left is concentrated around NLR (Netherlands Aerospace Centre).
Focusing on corporates
We can further filter the data by focusing on projects in which large, well-known corporates listed on the Dutch stock exchange such as Philips, Shell and Unilever are participating. This allows us to visually determine with which science and governmental institutions these corporates have co-innovated, and which projects they worked on together:
The visual shows that many corporates are indeed active in top sector project collaborations. The red cluster at the top shows activity of Unilever, FrieslandCampina, DSM, Wageningen University and DLO, and can mostly be related to innovation in foods and life sciences. At the bottom left in gray, we see significant activity from Philips, ASML and NXP, which all co-participated on projects with the Technical University of Eindhoven. The purple group at the center of the network concentrates on energy-related projects and collaborations, with Shell and AkzoNobel participating on various projects with ECN, the Energy Research Centre of the Netherlands.
Making public-private partnerships tangible
Even though partnerships between government, industry and science are often mentioned as key to a well-functioning innovation system, they are not often analyzed and visualized to provide a tangible view on the actual collaborations which have taken place. The analysis presented here provides an example of what is possible using publicly available data. Of course, many other visual perspectives are possible using the same data: for example, looking at sector-specific visualizations, the creation of participant-to-participant collaboration networks, a deeper view on project contents and keywords using natural language processing, et cetera. I hope this post provided you with an initial view of the possibilities – for any questions or feedback, you can leave a comment below or reach me on Twitter.
*: Using the keyword strategy for initial filtering may have left out public actors whose descriptions do not conform to the keywords used. We’re however confident that the bulk of the public organisations is included using this strategy.