Named after the Czech word “výzkum”, for “research”. I like to think mysterious Czech scientists pronounce it with a deep, dramatic intonation, like “Khazad-dûm”.

Vyzkumap is a hybrid search engine frontend and graph visualiser, used for exploring the huge Mendeley database of research papers. The basic browsing is functionally similar to that of the Mendeley website itself, but with a cool spiffy hipstery animated HTML5 pushState graph-viz interface.

I started development of this before Map My Field, but the implementation proved a fair bit more challenging. My original conception was based on hard citations rather than the more vague “related research” API command. However, acquiring the citations soon proved an intractable task (let alone quickly and reliably!). I also wanted to color-code by open access status, which the API presently doesn’t provide a means to bulk-fetch. So I spent a bit of time dithering about the speed-completeness tradeoff, eventually coming to a decent compromise by using journal names instead of individual paper ids. Aggressive caching is my friend.

So those are my two entries for the Binary Battle competition! I’ll be most pleased if either of them get into the top 10; I’m not really expecting to win (though I could certainly use the money). We shall find out shortly!

Because, sadly, real life does not afford me the opportunity to cover the planet in rainbow-coloured star stickers that expand into giant blobs.

Map My Field overlays the locations of Mendeley users in each academic discipline on top of a Google Maps instance. This may approximate the geographic distribution of research activity as a whole– but the usual sampling bias caveats apply. I don’t know how the probability of academics creating a Mendeley account varies by country, but I’d guess language barriers play at least a minor role. As of the time of writing, Mendeley has a total of 90381 users. It’s a decent number but not yet enough to include a truly representative sample in the smaller fields. For example, there is exactly one (Danish!) Queer Studies guy on the map.

From a technical standpoint, location density maps are hardly a novel idea, but the particular implementation details might be of interest. I use a transparent Raphaël canvas positioned over the Google Maps interface to draw the markers, which allows for all the extraneous fancy animations. When a node is clicked, the specific user information for that location is retrieved via AJAX and displayed in a dynamically positioned/resized HTML div to give the illusion of being “inside” the marker. Geocoding is pre-generated by a Rake task using the lovely API Google provides.

This is actually my second entry for the Mendeley API competition. I’m still refining the primary one; expect a post about that soon!