Figure 1
The flowchart is organized into three large rectangular sections representing the stages of data handling. The top section is “Data Collection”, the middle section is “Data Processing”, and the bottom section is “Data Analysis”. In the top section titled “Data Collection”, a box labeled “Web scarping: «coworking»” appears on the left. A rightward arrow points from this box to a cylinder icon labeled “Large Twitter datasets”. In the middle section titled “Data Processing”, a dashed arrow leads rightward from the “Large Twitter datasets” cylinder to a box labeled “Converting all text to lowercase”. A rightward arrow points to the next box labeled “Removing punctuation, U R Ls, hashtags, mentions, reserved words, emojis, and smileys”. Another rightward arrow points to a third box labeled “Implementing stop-words removal, tokenization, stemming, and lemmatization”. A final rightward arrow points to a cylinder icon labeled “Twitter datasets cleaned”. In the bottom section titled “Data Analysis”, a dashed arrow leads downward from the “Twitter datasets cleaned” cylinder to a box labeled “Identify the optimum topics number”. A rightward arrow points to “L D A: topic generating”. From this box, one rightward arrow points to “Labeling topics”, and one downward arrow points to “Extraction of sample data from each topic (manual check)”. From the manual check box, an upward diagonal arrow points to “Labeling topics”. A rightward arrow from “Labeling topics” points to a final cylinder icon labeled “Twitter Datasets labeled”. Below the final cylinder, a large downward block arrow points to the note: “Using in stage 3: developing a grounded theoretical model”.Words (COVID-19 related) frequency. Authors’ own creation