Professor Fouad Mrad, Frontier Technologies Manager at the United Nations Economic and Social Commission for Western Asia, and a judge on QF’s innovation TV show Stars of Science, on what data can tell us about sentiments surrounding the impact of the pandemic
As COVID-19 continues to accelerate, and the response to its impact must be faster than the problem, decision-makers need estimated evidence to guide their policies and their actions.
Big Data and Artificial Intelligence tools can support in providing an overview of estimated sentiments among communities in “real time”, to support nations in more adequately calibrating their policies and their response
Since evidence based on conventional surveys is time consuming – the typical cycle takes two years - and impossible with social distancing measures, Big Data and Artificial Intelligence tools can support in providing an overview of estimated sentiments among communities in “real time”, to support nations in more adequately calibrating their policies and their response.
At the United Nations Economic and Social Commission for West Asia, I conducted an exercise with Data Pop Alliance that utilized digital Arabic content analytics tools from Qatar Computing Research Institute [part of Qatar Foundation member Hamad Bin Khalifa University] to analyze, in several Arab countries, public opinion sentiments surrounding COVID-19 in relation to the health and education sectors.
It was decided to assess these public opinion sentiments through analyzing opinions expressed through Arabic tweets and posted media articles in these Arab countries: Lebanon, Iraq, and Kuwait. It involved the use of an open source Twitter scraper was used, which allows tweets to be fetched by keywords, locations, language, and dates by building a ‘query statement’ - specific information within a database - which is then requested from Twitter Search API.
Once this fetched content is gathered, the Arabic Sentiment Analyzer tool supplied by Qatar Computing Research Institute was used to analyze the sentiment in Arabic text as positive, negative, or neutral.
What this tells us is that the impact of COVID-19 on Arab countries is not uniform, and this shocking quake exposed many faultlines in their social, economic, and infrastructural fabric
After agreeing on selected querying keywords related to countries’ responses to COVID-19 on March 24, we found many of the sentiments within the 5,586 tweets scraped were neutral. In Lebanon, 47 percent were negative and five percent positive; in Iraq, 33 percent were negative and five percent positive; and in Kuwait, negative sentiments were seen in 47 percent of tweets, with positive sentiments expressed in 10 percent.
The public Arabic tweets that contained keywords related to education, universities, health, hospitals, coronavirus, and similar words gathered over 1,000 items in Kuwait and around 2,400 tweets in Iraq. While the percentage of negative sentiment did not pass 50 percent in any of these countries, it is noticeable that positive sentiments did not pass 10 percent, and the majority of the content was classified as neutral.
Of course, this is a sample batch, reflecting just one day. However, the trend over several days or weeks would give an indirect estimate of the public mood.
In addition to Twitter content, a live monitoring tool by Google GDELT was used to monitor posted media sources in the same countries during the same week in March. GDELT sentiment scores an analog number between (-1 negative) and ( +1 positive) for English and French media articles. A similar tool, Mazajak, gives a sentiment percentage for Arabic media content.
Focusing on the same keywords used in the Twitter analysis and the related themes – health and education – this monitoring exercise showed:
As the custodian of valuable data, governments can harness the available sources, along with open data, for feeling the sentiments being expressed through public opinion in terms of reactions to medical matters, security issues, social distancing adherence, and the socio-economic impact of the various responsive policies that are developed. This matter, as the pandemic is continuing to ravage societies, and plans for the recovery phase need to be in place quickly.