From the trenches: real data science project from start to finish | Hacker News Reader

From the trenches: real data science project from start to finish

datasciencecentral.com

32 pointsColinWright11 comments

Not really sure why this is getting up-votes.

Maybe I'm missing something but this article feels fairly incoherent and the analysis looks like something I'd expect your average marketing or product manager to be able to do rather than anything more sophisticated.

If you look at the other book this guy had published, it sounds like he has the same issues there: http://www.amazon.com/Data-Science-Analyticbridge-Vincent-Gr...

Sadly communicating one's findings as a data scientist is one of the most overlooked skills.

That's because that's all "data science" is. A buzzword (phrase?) used to make people who like the sexy sound of "big data" but don't want to put in the effort to become computer scientists or exert the time and effort to learn technical skills beyond whatever trendy statistics software they can find.

I agree, my first thought reading this article was, "this is a perfect example of why communication skills are just as important as technical ability".

Also, what was that surface mesh from the beginning of the article? There was no caption, no title, and no axis labels.

It would be nice if instead of simply saying "Well, this is rubbish" you could actually suggest alternatives that are of more value. Criticism is easy, producing excellent written work is rather more difficult.

So can you suggest better alternatives?

What are you looking for ?

This article was written for data scientists, not for investors. Here I try to explain how a data scientist (entrepreneur) can make a 7-digit revenue working from home, with no funding other than re-investing some of the profits into her business, with no phone calls, no travels, no meetings, based on a real story from a guy (me) born abroad in poverty. If it was that easy, everybody would do it. So I'm not surprised many people don't understand: if you don't understand, you can't do it - plain and simple.

I appreciate the value of civil discussion, but five minutes into scanning this piece I am unable to determine if there is any value. If someone has put the time into it, can you provide a useful summary? Or at least what does the multicolor plot represent?

They found that surface analysis indicates that google ads are not a good investment. But when they looked into what the users from google ads did after the initial page view, they actually signed up at a higher rate and thus the ads were worth the cost.

Awesome, thanks for that I'll give it another read.

is there a reason the charts are so difficult to read? especially after 50 posts a day on data visualization here on HN?

Reply on news.ycombinator.com