RSS 2014 Significance Lecture - The Big Data trap

[@DzyLabram]

I was a little cautious because of the title, after watching this video I think it's wrongly named. Maybe I just got to a different conclusion, I don't think Big Data is a problem, I just think that it is really important to think about biases.

[@philbx1]

Glad to watch this video. There is how the world will fail for a long time I think in that people rely on computers way too much. I think the 'optimisation' of everything in society is a very bad idea and the 'bottom line' of data isn't the best practice where humans are concerned. Not that average people will complain, as they trust the same computations as major companies and governments do.

[@a1akapokm]

Great talk, thanks for uploading

[@acexzavier897]

instablaster...

[@vwabi]

I think people are missing the point. In my opinion, this talk is not about why big data doesn't work. Of course big data can work, there are many succesful examples and data science generates billions in value for businesses. But we should try and think less of data science as "take as much data as possible and use machine learning on it". and more of it as "basically just statistics with some business stuff added on." There are rules, and following and understanding them has value. Like he says, many lessons of statistics have been known and are obvious for decades now. By ignoring them and just throwing everything into your sklearn model, you are not really doing proper science.