Facebook Research
labs.rs
labs.rs
On an unrelated note, how safe does facebook feel if they're advertising this?
> “Facebook will market you your future before you’ve even gotten there, they’ll use predictive algorithms to figure out what’s your likely future and then try to make that even more likely. They’ll get better at programming you – they’ll reduce your spontaneity. “ Douglas Rushkoff
Edit: After looking a bit more, is this put out by a third party? That would make more sense.
The first plot we hit is a graph with 10 node types and 4 edge types, IMO the decision to keep the entire post black and white looks good at first pass, but it starts to hurt on this graph. I am having trouble extracting any useful information from this graph.
Out standing questions on the first graph:
How does it differ from some other tech million/billionaire?
Why not collapse the cases where a node has only one parent and many siblings of the same type? (e.g. collapse all of Peter Thiel's companies, sans asana, into one hypernode)
Why is previous given the same weight as current? Why are titles not treated differently? (i.e. founder vs security tester (previous))
All told it's a very busy plot that I find really hard to derive any insights from. In fact much of the text that follows under the heading "Here are few examples of how you can read this map" contains information that is only tangentially referenced in the plot.
The second plot, under the heading "Management ties and education" is close to useless IMO.
The first half of it is a fairly typical representation of a bipartite graph, but has so many crossing lines it's near impossible to use. It seems to only be useful to look at the number of edges on one node.
The second half seems to largely just say that many of the powerful people in management are from top universities and disproportionally educated in the US. I love sanky diagrams, but flowing from university -> top 100 -> country feels like a strange choice to me, it erases the information of which university is in which country, but then again the majority of the chart falls into equal portions after cutting out Stanford, Harvard and Columbia.
I'd prefer a color (or texture) coded map with dots sized proportional to attendance, to this chart.
Anyways I should get back to work, I'm interested in going back over the rest of it. The authors have done a fabulous job at making these visualizations mesh well with the article and try to tell a story through heavy use of plots. I actually really like it. I'd love to see the code/data provided at the end. It'd be a fun exercise to try and improve upon them.
Also: Those are probably the worst visualisations I ever encountered.
There are so many terrifying layers to this headline