The Real Life of a Data Scientist
It’s obvious that data is making major headways in terms of its role in our lives. That’s good news for data management workers, but as I discussed in my previous post, the push to become data-driven also raises some serious questions about our ability to use data in responsible, appropriate ways.
You may not think that’s IT’s problem, but I disagree. As decisions become more data-driven, I think data modelers, data managers, CIOs and other IT data workers have a professional and perhaps moral obligation to help guide its use, at least in terms of insuring that the findings remain valid.
Frankly, I’m worried that data illiteracy might be a major barrier to embracing data-driven leadership.
As a world, we don’t even have enough data scientists to meet the job demand within four years. That’s pretty sad, when you think about how long we’ve been building databases and collecting large data sets. We’ve long known math skills aren’t where they should be, and I’m 95 percent sure math is a key part of analyzing data.
Also, statistics is a very small part of most high school math programs, and we already have a math literacy problem in the U.S., which is going to bleed into how people understand and interpret data. When it is taught, the focus is on the math of creating the statistics, rather than discussing how data can be manipulated or misused.
Even many college programs require only one statistics course—and in my case, I could take logic instead (a discipline that actually helps a lot with what constitutes a reasonable use of data, I think).
This should give data people pause: If business users are data illiterate, what might be the consequences of giving them more access to data and analytics tools without more education about how to use data?
I’ve already shared recent articles that demonstrate algorithms can outperform human judgment — but generally, we refuse to accept that and often go against what the data suggests. How much of that is ego and how much of that is ignorance about data and analytics?
Then, the question becomes whether every job lends itself to being data-driven in the first place. According to teacher Elizabeth A. Natale in The Courant:
Unlike my engineer husband who runs tests to rate the functionality of instruments, I cannot assess students by plugging them into a computer. They are not machines. They are humans who are not fazed by a D but are undone when their goldfish dies, who struggle with composing a coherent paragraph but draw brilliantly, who read on a third-grade level but generously hold the door for others.
