“Most of the research can’t be done by universities because we can’t access the data”

Arthur Charpentier was invited to teach the Summer School course on Big Data for Economics at the UB School of Economics. The course took place in Barcelona from Monday 2nd to Friday 6th July 2018. We had the chance to interview him after one of his lectures, which gathered 34 students from 14 different nationalities. Charpentier is a Full Professor at the University of Rennes 1. He also runs the blog Freakonometrics, with around 2 million visits per year and a Twitter account with the same name, with over 22.000 followers.

What implications does the data revolution bring for the field of Economics?

With social media you get information about what people are talking about. For example, you can extract the information from Twitter. It has already been used a little bit in finance, but can you predict that a stock will fall because a lot of people talk bad about it on Twitter? Some people claim it is working but I am not sure because it is complicated to know if something is positive or negative. Sometimes you get some kind of irony and it is very difficult for a machine to understand it.

What do you think about the kind of information that smartphones provide?

The smartphone is different because with them you can get the location. I think the location is interesting because you know exactly where people are. Especially if you want to design a city, it is good to know where people live and where they want to go, and I think that is interesting.

Does big data allow researchers to answer economic questions which could not be answered before?

Possibly. I remember when all the national institutes of statistics were trying to get the price index. They were surveying people, taking notes and so on. But now it is very efficient. However, it is not coming from the internet. It is coming from the machines that people are using in the market: you can get a lot of price information and see exactly when prices are changing, so it is improving a lot. Therefore, you really need less survey, you need less people asking questions. I think it is a good way to avoid very costly studies.

Has big data raised new questions in the field of Economics?

There are no new questions so far, but I think that big data is a good way to get an answer to all questions.

Speaking about the potential big data offers for economists, what are the limitations we still face?

There will be limitations because of regulations. If you think of what is going on in Europe, it will be more difficult to get access to data. I think it is more a question of setting rules saying what could be done and what could not be done. Now people start wondering “Can we really do that?” I think we have been a little bit too extremist. When I now ask for data, a lot of people say “No, I can’t give you that because we have anonymity rules”. I think we need to make clear that we do not care about the name of the person. I need the name just because I want to match it with another data set. I need it but I do not use it. It is just for research.

The scariest fact is that most of data is not collected by public institutions. Therefore, most of the research cannot be done by universities because we cannot access the data. Most of the research can be done by companies. They have much more information from Facebook and Google that we could ever have in academia. We cannot control what they have, which was not the case before.

Do you think that all economics should have a background on big data?

I think that a team with one or two researchers with a strong background in big data can still work with non-data related economists. The thing is to get a good mix. However, I am still surprised not to see a lot of economic faculties hiring experts on big data. I understand it is complicated because you need people with a good background in computational science, economics and statistics. Moreover, most of the good people will be hired by the industry.

 And about your blog Freakonometrics – when did you start it and why?

I started it more than 10 years ago, when I was hired for the first time as a professor. The university said that every researcher had to have a blog. My challenge was to proof that it is completely stupid. I said “Okay, I will start a blog and you will see it will fail”. But actually it did not fail. I made the blog mainly for my students. I was blogging a lot of material for them and then I got questions from people who were not my students. And I got questions from everywhere. It was helping me to improve the course. I do not know how to measure how well it is working. There are not a lot of comments, but there are a lot of visits and a lot of people talk to me about it. Indeed it is working well.

---