Librarians as data scientists. What should we be able to do?

Data is the next big thing. Big data is even bigger. And as always when something new happens, libraries jump on the bandwagon from the beginning.

We are coming to it at bit slow this time, at least in Denmark. We might have been intimidated by the Big in big data. But we’re getting there. And as always when we do something new, we have the problem of figuring out exactly what we should be able to do. Wich competencies should we have, what is the level of support that we should provide to our patrons?

I imagine that other places trying to get into data science, would begin by hiring a data scientist. Or at least send the most qualified personel on courses and training sessions. Not so in a library. We have all the books, we’ll just read them, and then we will know how to do it. It actually appeals to me as an engineer:

BOSS: We should do some data science
ME: OK, do we have anyone on staff that knows anything about that?
ME: Hmmm. Are we going to hire someone who knows anything about data science?
ME: Are we going to invest in training or courses to learn about it?
ME: OK, lets get started.

It gives us a chance to play around with exciting and difficult stuff, without being burdened with any actural knowledge about what we are doing. That is more or less how engineers define fun!

So. What should we be able to do?

  1. We are not going to be able to advise our patrons about what statistical test to use. There are simply to many. But we should be able to ask qualified questions: “Are you sure that that data follows a normal distribution?”
  2. We are not going to be able to advise our patrons about what data, and what correlations are interesting. We have no idea whether a correlation between a biomarker and a disease is relevant. But we should be able to recognise a correlation. Take a look at this page – its hilarious.
  3. We are not going to know all the intricacies of all the software we provide access to. Theres a reason we have 79 books in our catalogue about ArcGis. But we should be able to perform a simple data analysis in each of them.
  4. And of course – we are going to provide the world class level of customer service and assistance, that make libraries famous. Now we’re just going to assist with data science.

Thats it. How are we going to get there? I’ll get back to that later – now I have a meeting on LibQual.