The South African Twittersphere during Covid19
By Zach Wolpe & Keagan Stokoe
We’re living in unprecedented times. Whilst we face severe financial & medical challenges — one ought not to neglect emotional & psychological burdens presented by the break of normality. Whilst limited, social media allows for a window into the human condition. Inspired by the emergence of a sense of unity & connection, here is a brief analysis of SA Twitter.
Lets measure how South Africa feels about Covid19
First, let’s ignore Covid & assess the general emotion in all tweets. We sampled the Twitter activity in South Africa by identifying the current major trends & thereafter sampling recent tweets belonging to each trend (approx 10'000 tweets in total).
Code available at the end.
We then visualize the most commonly occurring word on the South African Twittersphere. Undoubtedly riddled with Covid references; glad to see ‘love’ at the top & what would twitter be without a football reference?
We can also extract more emotive words, a key component in modeling tone.
Lastly, we can approximate the sentiment on Twitter, which has been overwhelmingly positive. The positivity & negativity indices dwarf the other attributes as they often considered aggregations of the same cluster (for example, most tweets that are categorized as displaying ‘trust’ would also fall into the ‘positive’ category).
We can thus infer that Trust, Joy & Anticipation are the dominant sentiments — indicative of the hopeful but nervous nature of the country.
Secondly, let's investigate Covid specific tweets. We scraped Twitter for 10'000 Covid related tweets in South Africa. This visualization is susceptible to the same deceptively large ‘positive’ & ‘negative’ assignments.
Trust & Fear are the dominant emotions
Finally, we assess the movement of two major trending hashtags in South Africa:
#CoronaVirusSouthAfrica
#WhenCoronaVirusIsOver
Conclusion
It appears the republic of Twitter South Africa is hopeful but anxious about the situation, perhaps we will revisit the analysis as time passes.
Full Project: https://github.com/ZachWolpe/SmarterThanCovid
