Project: The scramble for narrative control - Unpacking Ethiopia’s civil war social media discourse
Date: February 2022
The scramble for narrative control - Unpacking Ethiopia’s civil war social media discourse
The scramble for narrative control - Unpacking Ethiopia’s civil war social media discourse
Earlier this year, we assembled and analysed a custom-built dataset of 88 million messages on social media across a two-year period, from February 2020 to February 2022 surrounding the ongoing civil war.
To track social media trends, we built a custom lexicon to reflect the context of conflict in Ethiopia and identified that Twitter has become an online ideological battlefield.
As of December 2022, this research has become even more topical as the harmful influence of social media is cast into the spotlight. This includes the involvement of Meta, as the company is now being sued, with claims alleging that its algorithm helped fuel the viral spread of hate and violence during the war.
Conversation volume
88.6 ml
Unique authors
3.06 ml
Total reach
7.1 bn
Applying narrative analytics
Through our unique Acoustics tool we identified the top narratives within the conflict conversation data set and focused on four priority narratives, selected from both sides of the conflict. These were; #Tigraygenocide, #nomore, #tplterroristgroup and #ethiopiaprevails.
Our research team also explored how imagery was deployed across the conversations to aid these narratives.
Our NLP-based emotions analysis identified the psychological states as the conflict unfolded.
Alongside tracing the evolution arcs of the conflict narratives, we conducted an in-depth NLP-based emotions analysis to identify psychological states and sentiment as the events unfolded.
Through mapping a spectrum of emotions including surprise, sadness, joy, fear, disgust and anger - our investigation revealed that sadness was the dominant emotion. We were able to pinpoint the causes and triggers of these emotions too.
What’s next?
Our data and insights can provide not only a retroactive overview of events, but also forward-looking key information for various decision-makers.
We are able to drill down to various geographies within Ethiopia to understand hyper-local sentiments, perceptions and narratives.
This conflict in Ethiopia has far-reaching implications for foreign policy and diplomacy across various countries like the US, China, Turkey as well as those within the region like Sudan, Somalia, Eritrea and Kenya.
Using this data, we can gain an in-depth understanding of the lived experiences of various stakeholders and their specific needs at real-time speeds.
We can channel other open source data to complement our social-data to bring in additional dimensions and bottom-up intelligence. For example, satellite imagery and socio-economic indicators.
There are a whole range of applications beyond conflict settings where our tools and methodology can be readily used. From impact evaluations to beneficiary needs assessments to new program development to policy design.