Social media played a significant role facilitating communication and interaction among *socially-aware* communities and participants of political protests.
People used social media to organize demonstrations (both pro- and anti-governmental), disseminate information about their activities, and raise local and global awareness of ongoing events
E.g. ↓
In recent years mainstream politicians strengthen their presence in social media platforms, recognizing the power of such tools.
## Our Objective
Quantify and present data on two italian politicians, **Matteo Renzi** and **Matteo Salvini** as a first step towards an appropriate study of the behaviour of influencial politicians online.
## Dataset
Our dataset consisted in 5370 posts by Matteo Renzi and Matteo Salvini publicly accessible on Facebook.
The dataset was obtained by scraping using [python requests](https://2.python-requests.org/en/master/).
## Main Difficulties
Scraping Facebook is a hard task if you can't request API keys.
* We got banned various times
* Page elements related to likes, interactions and responding users always change `id` and `class`
For these reasons we limited the analysis to just two politicians.
* page created in September 2009
* 959 posts from 01-01-2018 to 24-11-2019
* Posts from 2009 to 2017 were removed by the owner of the page
* In total 79 000 € were spent on advertising the page on Facebook
* page created in February 2013
* 4411 posts from 28-02-2013 to 24-11-2019
* In total 164 000 € were spent on advertising the page on Facebook
## Future improvements
Consider that a post could provide different values such as:
* increase partecipation
* provide information
* grab or detract attention
## Future improvements
Focus on semantic analysis:
* geographical data
* number of likes, shares
* rate of engagement
* sentiment analysis
* fake news detection
## Network Perspective
Model the analysis from an ego network:
* profiles of people following a politician
* spread of relevant/fake information
* model engagement and cascading behaviour with other users
* model social influence
* understand esternalities on a macro scale
Hong Kong protest in 2019