课程: The Non-Technical Skills of Effective Data Scientists

今天就学习课程吧!

今天就开通帐号,24,600 门业界名师课程任您挑!

Curiosity

Curiosity

- Years ago, I was on a project for a big national bank. And we were looking at bank activity. We found that a small but non trivial number of bank customers were checking their bank balance in excess of a 1,000 times a month. So we called one of the data experts on the clients team. He dismissed our curiosity is worrying too much. He suggested that we simply delete those client IDs from the analysis. Now beyond the fact that any cutoff would be arbitrary, why a 1,000 and not 998 for instance, we thought the explanation might be an app, like mint.com, or Credit Karma, or something like that. Again, we were accused of worrying too much, sorry, but this is simply unacceptable, and a good data scientist. Now that wasn't his role. So he gets a pass, but we had to proceed with or without his help to address the problem. Without direct access to the data, we had to get a bit creative. I ran across another striking example of this when I was visiting a cell phone company overseas. I asked…

内容