The Three Sexy Skills of Data Geeks | Hacker News Reader

The Three Sexy Skills of Data Geeks

dataspora.com

88 pointsdhotson12 comments

The hard part would be to convince a business that these skills provide value. It can be a large benefit to a company to effectively leverage its own data, and having someone that can not only extract and interpret, but also put the right visualizations in the hands of the decision-makers would be a great benefit to help a company, it can be hard to get those that hire to buy into the soft benefits. Its usually a major obstacle in any business intelligence project.

Web based companies already hire these people. Google, Paypal, Amazon, even "small" startups. I interviewed for a number of these positions in Silicon Valley last year but decided to do something else. Also, I don't really have the right stats background. Most smart businesses already know they have a data problem.

FWIW, the first two: stats and data munging are the most important. There are very few people who really know stats and can also set up a pipeline for ripping data out of log files and databases and get it into an analyzable format. The cool viz stuff, like what Ben Fry does is not so important. You just need to be able to make a couple kinds of graphs and charts.

+1 for bringing up the hard part. I've worked on several big data warehouse projects where it was tough to make business managers understand that even the simplest correlations between different fact types across dimensions (using those terms in the Kimball sense) provided ROI.

Forget visualization and deep statistics usage. You simply can't get there until you have executive commitment and belief in the value proposition. Personally, I wish I was a better evangelist for these skills.

I read about several management consultancies that offer such a quantitative approach for lots of money. Perhaps its easier to sell from the outside?

The point on statistics is well made. I recall my cryptography professor calling it "Sadistics." Still, I continue to read more & more on it, as it just keeps getting more and more useful.

Performance measurement of largish applications often comes down to statistical questions.

"There are three kinds of lies: lies, damned lies, and statistics."

- Benjamin Disraeli

all the more reason to understand it...

I do not think the "sexy" they refer to is the same kind of "sexy" that a geek would want to improve in.

I know I was disappointed that it wasn't a list of common geek skills that people are secretly drawn and attracted to.

Helping people share your geeky pleasure is indeed a sexy skill, especially if they're a little less geeky than you. Everybody wants to understand. Basically, if you make people feel a little smarter, they'll love you. Data visualization has that power.

I've never heard anyone call XSLT elegant before.

he was sarcastic in the post

fail to reject h0.

Reply on news.ycombinator.com