Hacker News Reader

Comment by marrone12 | Hacker News Reader

This is a cool idea, but I wish everything wasn't so 'big data' oriented. Most people will never work with big data. Instead of teaching me map/reduce, how about teaching me how to model with a mixture distribution? Teach me how to master small data and then scale those up to big data when and if need be.

View on HN

Cloudera seems to be sponsoring this class and both instructors are from there. I think that explains why the class has a "Big Data in Hadoop" theme, since that's what Cloudera does.

(as a comic aside to the buzzwordiness of "Big data", I saw a tweet yesterday that lamented rising use of the term "Hyper data". Which was followed by a reply about approaching "Ludicrous data", with a link to Spaceballs' ludicrous speed scene)

You might like Data Analysis from Coursera.

https://www.coursera.org/course/dataanalysis

From the post it looks like Udacity too is working on courses that address this.

I'm working through this course (and took the earlier sister course Computation for Data Analysis with R). It's quite good so far. They stay away from the "big" part and focus on the core of data analysis: how to find data, clean it up, explore it, find relationships and present your findings. We are using R, which is suitable for most data sizes. It's offered by Johns Hopkins, and has more of an academic bent than an industry one. Great general purpose knowledge that I think you would want before you start messing around with Hadoop.

I would also like something like this. I deal with data in my job all the time, everything from surveys to web analytics to customer data, but I only have a foggy notion how to work with it. As a result, most of the output is fairly simple, obvious metrics, but it would be nice to get a few online courses on how to do basic stats, construct surveys, do UI and A/B testing, etc.

Reply on news.ycombinator.com