Month: March 2015

Rob Gonsalves paints beautiful images where one viewpoint merges seamlessly into another. Often the transition is between a natural landscape and some human activity. The ‘human’, though, comes in the form of endless ritualistic repetition, so that there is somethign uncanny about them even in the areas where they are obviously human.

Here is more of his work

I’ve seen some truly awful infographics of corporate ownership structures. I’ve even occasionally perpetrated them. But this image is a classic of the genre:

Pretty convoluted, huh?

It’s from Muddy Waters, a much-feted research and short-selling firm. They are arguing that French conglomerate Bolloré owns a lot of itself through intermediate companies. We’re looking here at Financière Moncey, an indirect subsidiary of Bolloré, and the point is to show “how complex the relationships actually are” among these structures.

Look a bit more closely, though, and you’ll see that most of the complexity is artificially added. The diagram is just the same structures repeated again and again and again.

Here it is again, with all the repetitions deleted:

Fair play to Muddy Waters for figuring out the ownership structure. That kind of structure is painful to make sense of, and it’s easy to miss the circular ownership.

But they do seem to be deliberately exaggerating the complexity. Presumably the point is to show that things are so complicated that only their analytic genius can make sense of it. It goes with some snarky digs at analysts in the report itself — “

BOL has likely been misunderstood because the complexity of its structure makes it infeasible to use Excel to estimate the percentage of circular ownership

“. Much as I enjoy their approach, I wish they could make their point without, well, muddying the waters.

Why is nobody using markov chains to generate music playlists?

Playing music on shuffle is shit, full of jarring transitions, incongruous switches of tempo and topic. I masssively prefer listening to a carefully-mixed playlist, either via spotify or youtube, or an old-fasioned radio station like Radio Paradise.

So why not build a semi-shuffle? Start with one song. Look for all the tracks that get put onto playlists immediately after it. Play one at random. Then repeat the process — follow your second track with a random selection from everything DJs decided to play after it.

This is basically a Markov Chain.

It’s an idea blindingly obvious, relatively simple to implement, and in a domain that must appeal to thousands of CS students. Markov chains are immensely popular for toy projects online, since they generate fun output for very little coding.

But poking around online, I’ve not been able to find anybody using markov chains for playlist generation.

Here’s one of the more surreal corners of industrial medicine: crab blood donors. Apparently the best way of detecting some bacterial contamination is to add some crab blood extract to it, and see if it clots. So 250,00 crabs a year are scooped up and part-drained of blood, before being released with just enough blood left that most of them will survive the process.

To make it even odder: crab blood is blue.

[Thanks to Julien for pointing me at this bizarre micro-industry]