Bringing open data to life and making it tangible through visualisations

Which members of the Grand Council are the most active, and with whom? Do MPs work together across the political divide? Do men tend to work with men and women with women? This blog post describes how visualisation can bring open data to life and make it tangible, using data on parliamentary interventions in the Grand Council of the canton of Thurgau.

I asked myself the above questions when I examined the data from the Grand Council of the canton of Thurgau as part of the open data course at the University of Bern. Raw data, in Excel tables, with incomprehensible numbers, signs and variable names, can sometimes seem dry and lifeless. In my case, there were two such tables: One with data on the members of the Grand Council and the other with data on interventions tabled and their signatories* since 2012. My mission: to bring data to life, to make it visible, tangible and explorable. This is made possible by interactive visualisations that offer intuitive and entertaining access to the data, enabling users to discover links and acquire new knowledge.

Finding the right form

Every dataset contains a multitude of stories, big and small. The story you want to tell determines the form the visualisation should take. Do you want to show changes in time or space? Is the emphasis on numbers or relationships? Which dimensions do I want to show and link together? Depending on the answer to these questions, a map, a Sankey diagram or a tree map may be the most appropriate. Websites such as Datavizproject can help you get an overview of the multitude of forms of visualisation and choose the most appropriate one.

The questions that concerned me revolved mainly around who submits the proposals together, with whom and how often. The form of visualisation I chose is therefore a network visualisation, in which the MPs are represented by circles and joint interventions by links between MPs.

Kneading the data

Anyone who has ever worked with data knows that each set of data requires a certain amount of work until it has the desired form and quality. Often, the data needs to be sieved and cleaned like ingredients in a bakery, then mixed and kneaded into a dough and finally shaped into the desired form. That’s what happened in this case. Take, for example, the names of MPs, which were to be used to link the two datasets of council members and tabled speeches. Unlike identification numbers, for example, names are not unique and immutable. There are nicknames, double names, misspellings and name changes. All these elements found their way into my data. For example, one MP appeared under three different names (Gina Rüetschi, Regina Rüetschi and Regina Ruetschi-Fischer). In order to establish a clear link between the signature on an intervention and an MP, the names first had to be standardised.

The correct code base is the key

Once the data has been cleaned and formatted, the next step is to create the visualisation. Those like me who have little or no experience of interactive visualisations need help to get started. Fortunately, there’s plenty on the Internet, provided you know where and how to look. The tool I wanted to work with is called D3.js and consists of a javascript framework for interactive web visualisations. Tutorials on the basics are available on Youtube , for example. But if you have a specific visualisation in mind, you will quickly reach your limits with purely didactic knowledge. The solution lies in code examples. On Observable , for example, you can find hundreds of visualisations with the corresponding source code. The trick is to find the visualisation that best matches your ideas and use it as a basic structure, which you can then shorten, transform and extend until your own visualisation takes shape. In my case, it took two attempts before I found the right starting point: a minimal example of a simple force-based network with a filter function.

Let yourself be surprised

Even if you have a clear idea and the right tool, things don’t always go according to plan. Data can sometimes surprise us by going against our intuitive expectations about its content and structure. Initially, I assumed that collaboration on a given:MP was limited to a relatively small circle of people. As the visualisation developed, it turned out that the members of the Grand Council were much more interconnected than I had assumed. In fact, they were so interconnected that the image looked more like a ball of wool than a network (Fig. 1). It was almost impossible to draw any conclusions from this visualisation. The solution was to go back to the essence of the questions I wanted to ask the visualisation: Which Grand Council networks work together on a regular basis ? To achieve this knowledge gain, I had to reduce the information content. In the final visualisation, only the links are now displayed by default when two MPs have signed at least three interventions together. The result is a clear picture of regular collaboration (Fig. 2).

Conclusions of the study

The visualisation we have created allows us to explore the data and draw our own conclusions in a playful way. We can now answer the questions we posed at the beginning:

1. Which members of the Grand Council submit the most interventions and to whom?

The number of interventions submitted by a member between 2012 and 2021 tends to increase with the number of years of service since 2012. Toni Kappeler (PG) has submitted the most speeches, with 51 in 10 years. Toni Kappeler (PSE) and Josef Gemperle (PDC) have worked together the most often, having tabled 11 interventions in 10 years (8 of which were on the subject of energy). This result is also part of the answer to the following question.

2. Is there cooperation between MPs across the political divide?

Yes, there is a great deal of cooperation between the parties. Some MEPs, such as Elisabeth Rickenbach (EPP), even work more often with MEPs from other parties than with their party colleagues. However, it is precisely in the context of this regular cooperation that intra-party networks emerge (Fig. 2).

3. Do men tend to work with men and women with women?

Of the 264 MPs who have served in the Grand Council between 2012 and 2021, two-thirds are men and one-third women. If we consider the gender of the MPs, we can see that the network can be divided into a part dominated by women and a part dominated by men (fig. 3).

Fig. 3: Coloured network by gender

Once again, this separation is partly explained by internal party networks. The network dominated by women is largely made up of SP MPs who, at 60%, have a much higher proportion of women than average.
A second reason for this visible separation may lie in the themes of the interventions tabled. For example, while men are more likely to speak out on transport issues, women are more likely to speak out on social issues.
If this has piqued your curiosity, you can discover many more stories by viewing them live.

Open Government data and visualisations

I hope I’ve been able to show that interactive visualisations are capable of transforming dry data into lively and exciting stories. Often, the usefulness of Open Government datasets on obscure and supposedly uninteresting topics can seem limited. Yet visualisations help us discover the stories contained within the data. Stories that give us a glimpse behind the scenes of politics, the economy and society, that create transparency, reveal connections and encourage new ideas. At the same time, each visualisation and application helps to identify potential problems with the data and thus improve its quality. Open public data is the basis for an informed, critical, innovative and interconnected society, and visualisations help to develop this potential.

This article was first published on Open Government Data in the canton of Thurgau. On these pages, you will find information on Open Government Data (OGD) in the canton of Thurgau, events, press articles and applications, as well as the OGD blog.