What Data Scientists really do? We ask them for you.

Here’s a new issue of our divulgative column with the theme “Ask a data scientist” consisting in brief interviews with our D2 Seminar speakers. We are asking our guest to explain what their research is about, as if we were extending the Q&A session of the webinars. We intend to promote a public understanding of science and to reach the readers who may be interested in knowing more about the different fields in which data are involved.

This time we’re asking a few questions to Luigi Brugnano.

Luigi Brugnano was our guest to the Seventh seminar D2 series, he’s professor of Numerical Analysis at the Department of Mathematics and Computer Science “Ulisse Dini” at University of Florence. Before he presented his research, he dedicated his seminar to Francesca Romana Nardi, who was a dear colleague and friend, part of our Steering group as deputy director, and who passed away last October.

As explained in the abstract of his work “Recent advances in bibliometric indexes and their implementation”, bibliometric indexes are used to evaluate scientific production, research groups performance, and impact of journal. Even if these indexes doesn’t enter into the merits of a specific research, they can provide an idea of the impact of a certain study among the scientific community. As professor Brugnano pointed out, the currently used indexes often have drawbacks and vary for different subjects of investigation. For this reason, he and his colleague professor Amodio proposes an alternative index, based on the Google PageRank method to rank web pages by random surfer model. PaperRank and AuthorRank are the new indexes proposed that has also been recently applied in the Scopus database.

As we were listening to his speech, some aspects aroused our interest, and here’s what we asked him.

Could you explain to someone who doesn’t know what a bibliometric index is, why it’s so important to have an evaluation method to measure the impact of a research or a study among the scientific community?

As I told in my talk, bibliometrics indexes are only a rough measure of the impact of research on the scientific community, so that they should be used only to have a gross idea of the impact of a research. However, for this purpose, they can be very useful.

How the alternative index your study propose can improve how we rank papers, authors, and scientific journals?

Many indexes have a lot of inconveniences and, moreover, they usually strongly depend on the specific field of research. A more standardized approach was then needed. The proposed index is based on the very same approach of the Google PageRank, and everyone knows that Google is very effective in providing a correct ranking of web-pages. The proposed index is based on a suitable simplification, only exploiting local data, in order to save complexity. Nevertheless, the final result appears to be quite close to the original Google ranking, so that we expect the new index to be effective as well.

Luigi Brugnano is also a member of our center. More informations about his research can be found on his personal page.

You can access to the recording of this and all other seminars at this link. (Registration needed)