I got an email recently asking something along these lines: I’m a smart ex-engineer who likes stats. I want to be a data scientist. How difficult will it be for me to find a job doing data science work at a startup? I think a lot of people have this question. It’s interesting that it’s common given how hard it is to find a quality data scientist. I sent back an email which looked more or less like the following post - these are my thoughts on how to get hired, as seen from the eyes of someone who actually really wants to find more talented data scientists to hire. Quantitative intuition The first thing that I’m looking for is general quantitative sense. This is a really low bar. It’s not really important whether someone knows the quirks of Cauchy distributions or whatever.……

Read more →

Web apps seem to graduate through a series of stages with regards to analytics. Stage 1: We should probably keep track of page views, you guys The “is this thing on?” phase. Most people start with Google Analytics or something like it. Nobody is looking for a lot of depth in the numbers - all you want to know is whether people are using your adorable little web app or reading your blog or whatever. Basically, this is about seeing whether you need to do a better job promoting yourself. There’s probably somebody in the company who compulsively checks on how many visitors you’ve got. Nobody really knows what they’re looking for, but 10,000-visitor days from being on the front of Hacker News get everyone excited. Stage 2: We should probably track signups, too, you guys The “okay we’re ready for the users now” phase.……

Read more →

You can find gobs of chatter out there about what makes a good data scientist. Most focus on technical aspects: know statistics, write scripts, build machine-learning tools. Sure, these are all important, but the list is incomplete. The most valuable skill for a data scientist is the ability to shape the culture around them, to convince others to use data to inform their decisions. The best data scientists are the ones who transform their company. They motivate an entire organization to make decisions by measurement and numbers. They do not do all analysis themselves - they turn those around them into analysts. People who have machine learning skills, know statistics, and can write code are terrific, yes. But those alone are not transformative. why do companies want data scientists? When companies hire data scientists, they rarely want a wizard who can dazzle with highly technical presentations but who works alone.……

Read more →