Introduction
Twitter helps politicians to directly communicate to the public and to disseminate their message timely and widely. In this regard, former US President Donald Trump was a well-known Twitter user making over 25,000 tweets during his presidency.
PM Abiy Ahmed is also a great user of social media. He has Twitter, Facebook and Linkedin account in his official name. On Twitter, he made about 4 tweets per week over the past year in the Amharic language alone.
Analyzing a politician's Twitter account could provide insights regarding the political discourse of the leader and the overall acceptance of his/her ideas by the larger audience. Consequently, I will try to address this question for Ethiopia’s PM, Abiy Ahmed Ali. We do this, by looking into which types of tweet terms dominate his tweet as well as the sentiment of the tweets. Moreover, we do apply the same type of analysis for the responses of 18 recent tweets to gauge the overall response of the PM’s audiences.
The data is downloaded using the twarc command line tool. To this end, Wordcloud and VADER sentiment packages were used to analyse the frequency of words and overall sentiment of the tweets and replies.
PM Tweets
Here, we cover about 1006 English and Amharic tweets made by PM Ahmed during the past one-year period. These tweets are largely dominated by national, regional, and continent-wide narratives. For instance, ‘Ethiopia’, ‘Ethiopian’, ‘national’, and ‘country’ are among the leading words included in his tweets. ‘Africa’ and ‘regional’ are also mentioned several times in the tweets.
This might not be a surprise given that ‘Medemer’, an Amharic term that means ‘synergy’ or ‘coming together’ has been a popular concept used by the PM as a guiding principle of his leadership. The ‘Medemer’ philosophy advocates the idea of togetherness for a common purpose.
The positive vibe in the PM’s tweets becomes more clear when we find terms such as “one”, “work”, “support”, “will”, “key” etc. as the second most prevalent words. In fact, the sentiment analysis of these tweets reveals that almost every statement is dominated by positive words in a consistent manner (Figure 2). In this regard, Figure 2, shows the number of positive, negative, and neutral tweets that each of the top words appeared in the previous cloud of words. In other words, the figure tells the sentiment of each of the tweets that consist of the words. For example, two of the word ‘region’ mentioned in the following tweet are labeled as a positive sentiment as the overall sentiment of the tweet is positive:
Replies
Quality of replies: before heading to explain the findings of the replies to PM Abiy, it is worth noting the potential presence of fake accounts that could misrepresent the true public response to PM’s tweets.
As you can see from the figure below, there is a sudden spike in the number of new followers since early October 2020 that coincides with the beginning of the crisis in Northern Ethiopia. This is a clear indication that fake accounts are being used either to counteract or promote what is being tweeted by the PM.
Top reply terms: Regardless of the above caveats, we looked into the top words found in the replies to the most recent tweets of PM Abiy Ahmed. Here, we find terms like ‘Ethiopia’, ‘Ethiopians’, and ‘country’ are mentioned once again in the replies to PM Abiy Ahmed’s tweets. We also find the wartorn region - the ‘Tigray’ region, and the upcoming ‘election’ mentioned in a noticeable fashion.
The overall sentiment of the replies: A closer look at the sentiment of each of the responses to PM Abiy Ahmed’s tweets confirms a mixed feeling as a significant proportion of the mentions found out to be negative. Despite the limitation in the presence of fake accounts, the angry responses from the audience might indicate the unmet demand for tailored information that the PM needs to address.
By Yared Hurisa, Applied Scientists and AI for Peace Adviser
Comentarios