It's actually because Twitter rate limits the oembed endpoint by IP, so the app caches the oembed codes and uses those if it gets limited. I've been meaning to flip that bit to the client but haven't gotten around to it yet.
This isn't a problem with guys or with developers in general.
The vast majority of people who participate in conference are good people. However, it only takes one person (out of thousands!) who has this kind of personality dysfunction to lead an attack, and women are much much more likely to be the target of their dysfunction.
It's a play on "duck duck goose", the children's game. The similarity to duckduckgo was unintentional, but I like it. I actually wanted to call it "bookroulette", but all permutations of that domain were taken.
To be clear, I'm not knowledgeable about online advertising. I do understand that the incentives of the advertisers and the users are often opposed, but I was highlighting this as an example of poor product design likely due to optimizing click-throughs, not product experience. :)
In most data analysis work, we assume that the data resides in some database and that you have the luxury of iterating over that data as many times as you like to get to a final result.
The challenge with stream analysis is that you are dealing with a continuous stream of data where you can see each element of the stream only one time and must still be able to cluster/classify/analyze it. There are still few algorithms and tools designed explicitly for that purpose.
Don't read too much into that example -- I chose it as a humorous metaphor, not a mathematical argument, and I messed up the delivery in the talk, anyway.
I started this job before "data scientist" was in common usage. My title was scientist. Then we hired more scientists, so I got a promotion. :)
We build systems on top of the bitly data. This involves things like algorithms for identifying spam and malware, real-time content recommendations (which you can see in http://news.me), categorical analysis, etc.
I'm trying to use the model of the unix command line -- each script is useful and simple on its own, but in combination they become incredibly powerful.
I agree. For example, why do we look at email ranked chronologically? I'd much rather rank it by relevance or importance (or my likelihood to actually answer it).
The feature set includes sender, recipients, whether I'm the only recipient, time of day, words in the subject, words in the body, and whether I've exchanged messages with the sender recently.