Hybrid Format

A finite mixture approach with Interval-Censored Data for the analysis of digital divide through Facebook data

Dalila Failli (University of Florence)
Department of Digital and Computational Demography, November 06, 2023

Hybrid Talk Seminar, November 6, 2023 from 11am to 12pm (CEST)

Abstract

The Facebook Marketing Application Programming Interface (API) is a tool that provides access to the Facebook Adverts Manager from which it is possible to derive the approximate number of Facebook users who match specific characteristics. Our aim is to analyse digital divide through the Facebook API data via a finite mixture approach for interval-censored count data. Indeed, in this framework, response variables represent the number of users expressing interest in a certain digital topic. The proposed model allows us to perform a clustering of users’ profiles, which are defined according to age, gender, education, country, and marital status, on the basis of the digital interests of their units, as well as to cluster digital topics according to the users’ basic attitude towards them. We expect not only to obtain homogeneous groups of Facebook users and digital topics, but also that the level of digitalization is strongly influenced by age, gender, and educational level. Furthermore, the proposed model should be able to account for the interval-censored count data of Facebook’s API. 

About the speaker

Dalila is a PhD student in Statistics at the University of Florence. She holds a BSc in Economics and a MSc in Statistics from the same University. Her research focuses on finite mixture models for model-based clustering of network data.