Relations Between Coronavirus (Covid-19) Pandemic Data and Social Media Interactions: Big Data Mining Implementation – Turkey 2020 Case Study

This study by İlayda Dilek and Mustafa Coşkun examines the relationship between COVID-19 pandemic data and social media interactions in Turkey throughout 2020 using big data mining. Analyzing public Twitter accounts following the official @saglikbakanligi account, the authors filtered an initial dataset of 209,887 followers down to 2,891 real, active accounts using statistical criteria based on establishment date, account privacy settings, biography length, and tweet frequency. A total of 2,642,347 tweets were collected and subjected to sentiment analysis using a Turkish emotion dictionary and sentiment algorithms. Pearson correlation analyses revealed statistically significant positive relationships between daily pandemic metrics—including patient numbers, death counts, and recovery cases—and social media interaction variables such as tweet volume, retweets, replies, mentions, and average tweet length. Conversely, increases in total patient and death statistics negatively correlated with average retweet rates and general user sentiment, reflecting public distress and reduced interaction during severe periods. Additionally, validity checks showed that significant national events, curfew announcements, holidays, and major news events triggered distinct peaks in user engagement and mood fluctuations. Ultimately, the findings demonstrate that social media big data mining is a highly effective, valid method for monitoring societal dynamics and public emotions during crises.