We wanted to study the correspondence between the election results of regional elections and the activity of candidates and users on social networks.
the objective of the research was to analyze Twitter activities related to the regional elections in Emilia Romagna and search for possible relationships with election results.
The model was then applied to three other regional elections, leading the analysis to the same conclusions: Lazio, Lombardy and Calabria.
The analysis was conducted through two main threads: candidate activity and user activity.
Lucia Borgonzoni, candidate for the center-right coalition, stood out for increased activity on Twitter throughout the campaign period.
Daily frequency of Tweets by candidate
The graph shows the daily activity of tweets posted by the candidates (the straight line represents the date of the election). It is easy to observe how Bonaccini was the first among the candidates to become active on the social network, followed by Borgonzoni and Collot who instead start their activity the following month. One can also see several peaks of activity on different dates for each candidate probably related to news events.
Cumulative Frequency of Tweets by Candidate
The next step was to use the cumulative frequencies to give an overall view of candidates’ activity. From the cumulative frequencies, it can be seen that although Borgonzoni began her Twitter activity the following month, she has a higher and more consistent trend of activity over time that has allowed her to surpass the other candidates in terms of tweets posted as early as about a month before the election date.
Nevertheless, the policy proposal did not convince the electoral base, which did not reward her in electoral terms.
Most used hashtags for candidate Borgonzoni
most used hashtags for candidate Bonaccini
most used hashtags for candidate Collot
The top 5 most used hashtags on the official pages of the 3 most influential candidates were analyzed: Bonaccini, Borgonzoni, and Collot (not all candidates in the election had a Twitter page; for example, Battaglia has not been active on Twitter since 2013). The graph shows by frequency the hashtags used by the candidates and allows us to compare the issues most felt by them. A first glance reveals the different approaches of the candidates: Borgonzoni uses more self-referential hashtags as opposed to the other two who more generically emphasize the territorial component.
What different communication strategies were used by the three main candidates? Through the graph we tried to give an answer to this question. Through the graph we tried to give an answer to this question. As shown above Borgonzoni is the candidate who posted the most on Twitter, however, it is Bonaccini the one who used images and photos in his posts most frequently (72% of his posts contain at least one photo). As for videos, the low usage can be noted.
Stefano Bonaccini is the candidate who used the most hashtags and the most photos in his tweets.
Among the 20,524 user tweets analyzed, some trends emerged that help to understand the behavior and perception of the electoral base.
#bonaccini is the second most used hashtag
#borgonzoni is only in fourth position
One of the key aspects of the analysis regarding the January 26, 2019 regional elections in Emilia Romagna comes through the analysis of hashtags used by users. The graph shows the different hashtags used with their relative frequency. As expected in the first places we find generic hashtags related to the theme of the elections (#ELEZIONIEMILIAROMAGNA), followed by the names of the main candidates and politicians. We also find references to Calabria’s regional elections as they occurred at the same time as those in Emilia Romagna.
Temporal distribution of tweets
Temporal distribution of tweets (LOG)
Another aspect analyzed and shown in the graph concerns the temporal distribution of the number of tweets posted regarding the elections. From the graph, there is a peak in the distribution of tweets covering the week of the election, with the highest peak occurring the day after the election date. To observe the time trend in more detail, the same analysis is also shown on a logarithmic scale.
Comparison of number of hashtags, users who posted tweets, and unique active users
Comparison of number of hashtags, users who posted tweets, and unique active users(LOG)
Comparison of number of hashtags, users who posted tweets, and unique active users (cumulative)
An interesting aspect that was analyzed is the comparison between the number of hashtags used, the number of users who posted tweets, and the number of unique users active on the social network. From this analysis, as shown in the graphs, the spike in activity close to the election date is immediately evident. It is also relevant to note that hashtags, users and unique users show the same temporal trend, and do not show significant scissors: this means that an increase in hashtags posted corresponds to an equal increase in users posting. This is even more evident if values are plotted on a logarithmic scale as shown in the first graph. Finally, the same trend is depicted using cumulative frequencies, from which the analyses conducted so far are confirmed.