applied to ~85 data science positions. I can't even get recruiters to call me for a phone screening, so don't feel down =)
1. Previous internship in data science 2. Experience developing R packages and putting them on GitHub 3. Really having statistical theory down pat
You don't need an advanced degree to be a data scientist but you need a strong understanding of stats and how to work with data. Having an advanced degree is a good indicator that you can do that. But it's not a prerequisite for an undergrad: Github, internships, TA-ships can make that up.
I think one advantage is that while PhD's are typically very good at the research process and the techniques used in their research, undergrads could be more flexible and adaptive to different situations.
Or maybe there's a better explanation for asking for a candidate with a Ph.D. in Statistics to create a linear regression model in Excel. Because truth is, for many companies that's all they need.
Linear regression, logistic regression and k-means clustering, if you can get a project into actual real-money production on one of those, you are already well ahead of 90% of data scientists. And these techniques are decades old!
But also recall, that there are alot of people with general computer backgrounds and a whole lot more real-world work experience who could pick up the basics of data scicnce stuff as well..
Places where these jobs tend to be needed are often very highly placed and very strategic, with the decisions being made based on the data very critical to the overall company health. Some of what is being requested by requiring this level of credentials is real-world work experience and a certain level of maturity/professionalism that is often implied by this as well..
If the level of skill is basic enough that any other regular employee could learn it in short order, why wouldn't they and then have the cushy 'sit around and play with numbers all day' job too.. There needs to be enough 'there there' to placate internal politics as well..
Further, having a more advanced person would also likely need to be taking leadership roles in picking and rolling out solutions with long term costs associated with them, training other employees, etc, which also comes with the professionalism/experience part..
Definitely will take note of that in the future though.
As a new grad that went through the hunt very recently - it was a messy process. Very few places will consider you without extensive experience, or a masters/Ph.D. Of course if you're hiring people to research machine learning algorithms that's justifiable, but plenty of the responsibilities people associate with data scientists don't require advanced degrees.
And the number of posts asking for 5, 7, even 10 years of experience... absolutely astounding.
As someone uninterested in going back to school, I've resigned myself to getting some work experience and doing personal projects for 1-2 yrs before trying again.