When analyzing the social media sphere in Belgium I always find it hard to differentiate the Eurobubble from the “real” Belgian environment.
Belgium has a very active and engaged online population, with users from many segments of society. In addition to politicians and journalists, many entrepreneurs, consultants and professors active on social media.
“The Belgian society sphere” and the “Eurobubble sphere”. Especially when mapping social media interactions we can see how these two dimensions overlap and interact quite often.
While this creates some challenges on the monitoring/reporting side (as it is hard to distinguish and measure interactions coming from the two social media spheres) it provides opportunities to penetrate two social online dimensions at the same time.
Belgium’s linguistic division is also clearly visible on social media. While the issues on which they focus often coincide, there are almost no platforms or influencers that operate bilingually in French and Dutch. Publishing in English as a second language is very common too.
While these two dimensions may at times overlap in mapping exercises, they rarely overlap when looking at type of content which is shared and discussed. The expats community is the real engine of information and content sharing about EU-related affairs while this seems to be a minor concern for the locals.
What do you think about it? Can the two dimensions be separated or will it be like breaking an atom?
17 thoughts on “Discerning the Belgian social media sphere from the Eurobubble: mission impossible?”
I’d look at this from another perspective. What questions do you need answered from either Belgian people or Eurobubble folk? By isolating those questions it is possible to drive into conversations that are by Belgians (for example) but bear in mind people like you (who are in the bubble) will also have relevant views on Belgian day-to-day life i.e. problems with STIB, Belgacom, a nice restaurant etc. I’d still want to listen to those views…From a technical standpoint we use conversation clusters and influence measures to listen to the “right” people that tends to be enough but maybe you have a deeper problem I don’t understand.
Thanks Simon, have you ever done a similar analysis with Brandwatch? The origin of my question comes from my (very limited) experience in monitoring and mapping EU topics in Europe. Most EU-related events, conferences, chats etc.. taking in place somewhere in Europe (apart from BXL) always trigger huge amounts of SM conversations in Belgium and in Brussels. What I’m trying to look into now is to understand how many of these euro-topics-related conversations are generated by the belgo-sphere (NL-FR and locals speaking EN) and how many, or in which volumes, these are discussed by the Eurobubble. It’s not gonna be easy, but I have a few criteria in mind that could help answer these questions (in the near future)
Ah ok, so the theme is European issues such as enlargement, single currency and say roaming but the conversations generated by “normal” Belgian citizens who aren’t in the European bubble. So you can analyse those conversations and compare to what the French or Spanish think about Europe as a whole. Real Belgians that talk about European issues. What issue, I’ll see if Brandwatch can do it.
As it often happens in other countries too, the EU is discussed only in the ‘extremes’. I do agree that Flemish, Walloon and BXL citizens have totally different preoccupations (partly because the media frame it this way). Only a minority communicates in both languages and translates the issues, problems and debates into two languages.
That’s a very good analysis Kwinten. The problem at the monitoring level is mostly related to mapping. When the EU is discussed in Belgium and in Brussels it’s hard to distinguish how many conversations are created in the belgobubble or in the eurobubble (at least on long time frames)
In social media just like in the real world the Belgian social media sphere and the EU people are two different communities, they share some ties but are distinct.
Conversations are more likely to occur within a community than between the two. Simon’s approach of clustering the conversations is therefore valid, you just need to find out who talks to who in your monitored conversations.
Another way is to analyze the actual social graph (= the set of follow connections between the whole population). This gives impressive results: e.g. we’re not only able to differentiate French-speaking from Dutch-speaking communities, but also within Brussels twitter users to identify subgroups (bloggers, geeks, politics, media crowds).
That sounds like a very interesting analysis to carry out. Finding the right criteria to filter out the Eurobubble from the belgobubble can be challenging but the availability of tools like NetLogo, Gephi or NodeXL can provide an answer. Thanks Mehdi
just to illustrate, here is a subset of the Belgian twitter social graph: http://twitpic.com/didrma
French-speaking and Dutch-speaking communities can clearly be distinguished, this is because there are significantly more follow ties within communities. Same could be done between Belgian and Euro socmed crowds, jsut I didn’t have the dataset ready on hand.
about the graph:
Each dot corresponds to a twitter account, and lines represent follow relationship. Layout is computed according to proximity of people :the ones who follow each other get close to each other.
It’s not only the social media sphere vs the eurobubble. This same gap can be seen in mainstream media in all member states. I don’t see the two dimensions getting closer. On the contrary. Because local problems are always prevailing over regional or global ones. But, really, a highly interesting topic. Worth deliberating further on it
Thanks Adelina, I hope to expand the discussion to other comms pros
Sorry Mehdi, but there is a huge mistake in your graph: Dutch-speakers are on the right (in green), French-speakers on the left.
And thanks Marco for this really interesting article. I personally find myself in-between your categories: Belgian, part of Eurobubble, following EU related influencers as well as French-speaking and Dutch-speaking influencers.
Oops, indeed. I usually know my right from my left, just typed too quickly without checking…
Nice post Marco. Although I have no data or mapping to reference, I have the impression that the Belgian and Eurobubble social media spheres are entirely separate and reflect the clear real world split between what happens in the bubble and what happens in Belgium. I am in both the bubble and Flanders on any given day and my conversation topics rarely overlap:)
Thanks for sharing your thoughts on this Brett. Do you think a separation between the belgobubble and the eurobubble applies more to Flanders, Wallonia or the Brussels-region? I suppose in the latter we can see more overlapping.
A semi-automated solution that comes to mind is to scrape the self-disclosed user bios and to search them by keyword. Quite a number of Eurobubble-affiliated users might refer to the Twitter handle of their organisation in their bio, I assume? That would at least give you a percentage for a random sample of tweets. However to make that information useful on the long term, you’ll probably have to go back into your listening solution and tag these users individually…
It’s an interesting idea/challenge thought!
You’re right. I thought about the Twitter bios test. Already, a short analysis with Followerwonk can give some pretty neat results in analysisng the bubble
Good question to raise Marco! As Brett says I don’t have any data to build my argument on either but my impression as being Danish (from the few Belgian people I follow on twitter) is that they keep the eu data separated from the Belgian stuff just like other nationalities working in EU separate local stuff from the EU bubble. But I’m not sure if this is correct, it’s just an impression.