Monday 25 November 2013

Two hops from FHS: What can our Facebook page tell us about our network?

By Jeff Knezovich, FHS Policy Influence and Research Uptake Manager, Institute of Development Studies

The other week I had the good fortune of participating in an excellent meeting in Prague hosted by the Open Society Foundations: Policy Research, Technology and Advocacy Event @ the Hub. The event was designed to bring experts together from across Central and Eastern European think tanks to share ideas and learn from each other on innovative approaches to evidence-based advocacy and communications.

There were a number of interesting sessions from the two days (which I've expanded upon elsewhere) -- but I particularly appreciated a hands on session led by Josef Slerka, a lecturer in new media at Charles University in Prague. He showed us how to use some freely available tools to gain a better understanding of our position within social networks.

Making a network map

Based on any given page on Facebook, he walked us through how to find and map a network of other 'pages' (not individuals!) that like or follow that page. We started by scraping information from Facebook using Netvizz. They allow for mapping of up to two hops from any Facebook page, so I thought I'd start with the Facebook page for Future Health Systems (but Netvizz can also help you create a map of your own personal network, which I also did -- it was interesting to see the relationships!).

Netvizz produces a GDF file designed to be integrated with the network mapping programme Gephi (open source and free to download – but if you’ve recently upgraded to Mavericks on your Mac you may need to re-install Java first).

But it effectively creates a CSV file (comma delimited), which I imported into Excel, tidied with the vlookup function (probably my favourite Excel function, if I'm honest!), and then imported the whole thinginto GoogleFusionTables. If you haven't used FusionTables before, I recommend highly that you give them a go -- they can help create maps and lots of other interactive charts, not just network maps. Here's the result:

Click through for an interactive version!

One of the nice things about the interactive version is that you can reduce the number of nodes that it shows so you can really get a clear sense of how connected various parts of the network are. Also, if you'd like to open up the hood/bonnet and see how simple this really is, you can see the full Fusion Table (it also allows you to filter the nodes to explore specific relationships).

What does this tell us?

So we've got this map, but what's it good for? I haven't sat down to do a massively detailed analysis of this network, but even cursory glances can start to tell us a lot about the organisations that are particularly connected in this network.

  • Follow the donor: It's probably not that surprising, but some of the biggest nodes in this network are funders -- DFID, the Bill and Melinda Gates Foundation, UNICEF and USAID certainly stand out.
  • Birds of a feather cluster together: It's interesting to see some of the 'outlying' clusters that come together. There's a little New York Times cluster in there, and Johns Hopkins cluster (not surprising considering that they are the lead FHS partner), and a massive UNICEF cluster.
  • Strategies for developing a following: It's not entirely fair, as this is network is only two hops from the FHS page (and so it doesn't paint a ful picture -- but it is possible to add to the network!), but it's interesting to see the patterns of who follows whom on Facebook and the different strategies that the organisations are employing. Some follow many, while others hardly follow other pages at all. I'd be interested to see how the relates to cross-sharing of information.
  • Media still matter: Two clear nodes from the traditional media stand out. The New York Times and the Economist. Time to strenghten those media relationships, I guess!
  • Two hops to hop: The Guardian did a great interactive graphic relating to the NSA leaks on how quickly a network expands from hop to hop. This network map already gives me a clear sense of how quickly that expands, but it also gives me some great ideas of other pages to follow that I didn't know existed to grow a closer network.
But all in all, it was great to get the practice in developing a network map. Several FHS projects have been experimenting with social network analysis (SNA) to the policy influence and research uptake planning in their interventions. Knowing how to use tools like this can hopefully help directly strengthen our programme implementation. We'll keep you posted!