Introduction
After the definition of the main actors involved, we decided to focus on the debate in the academic field. This was mainly motivated by two reason: the fact that both from the gephi map, than from the Sven analysis we encountered many terms related to scientific, psychological and technical field, and the second reason was the fact that Google itself suggests to look for the academic results when we searched for our queries.
So we focused on two main scientific database: Google Scholar and Scopus.
We tried to use the same queries that we used on the second protocol, we filtered them by relevance reading the abstracts and, when possible, introductions and conclusion, and we collected some information about every single paper (name, author, year, university, country of publication, position in the debate, category, type of paper and number of citations), for a total of 63 paper.
We decided to expand the number of paper by considering also all the papers that cited the book “The Filter Bubble”, after having verified the central role of the book in the debate. The total of paper was 109.
In this graph we can see the different partition of the opinion for every typology. As we can see the first thing to notice is that in privacy the pro position is totally gone, no paper talks about web personalization in a positive way about the privacy. Another interesting fact is that in society and information the contrarious are more than the neutral.
To better understand where the academic debate is leaning, we decided to try to expand the number of analyzed papers. We started from the one that we considered in the last section and we decided to try to map them by starting from the one with a much more defined position in the debate.
The highlighted dots define the papers with most explicit position about our research question, while all the grey dots are all the rest of the papers, connected in relation of how many times they cite themselves. It is possible to see where all these papers are positioned in the debate, like for example we see that there are much more papers in the middle-moderate area.
This chart has been realized with a overall corpus of 3028 papers, obtained from the 63 papers that we initially analyzed from the queries. We created an excel file with two columns, source and target. In the first one we pasted every one of the 63 papers as many times as they were cited on Scholar of Scopus, and in the second column we pasted the citing documents. Once completed this file we opened it in Gephi and we locked on four different positions on the opposite sides the paper with the most radical and moderated opinion about the web personalization. We paid attention to balance their respective weight (in term of connection) in order to avoid mistakes in the forces involved. Then we decided to clear all the nodes that had just one connection with another node (1 degree), rank their dimensions in relation to their degree range and finally let Gephi map all the remaining nodes.
It’s evident that the great number of papers is collocated in the moderated area of the chart, lightly leaning to the PRO section. This is probably due to the fact that there are many papers which belong to the much more technological/neutral category.
