Time Series Analysis on Covid 19 Summarized Twitter Data Using Modified TextRank

ABSTRACT

To interpret people’s sentiments using Covid 19 twitter data, we discovered that people’s sentiments are distributed across multiple dimensions. Six significant themes were chosen, including administration, disease, healthcare, location, precaution and citizens, because these topics receive a lot of attention on social media. In this paper, we used the modified text rank extractive summarization approach on Covid 19 twitter data to reduce data volume without sacrificing the quality. The keywords are chosen from the pre-processed data set where the frequency of the words exceeds a pre-determined value decided through several trial runs. The goal is to extract the most number of word sets possible from the original tweets. These keywords have been grouped into the six categories listed above. To understand the trend of the topics, all the keywords belonging to a specific topic is searched in the summarised file of a given day, the count of that topic is increased on every successful match of the search. The graphs for the counts of all the themes for each day were then plotted. To identify patterns, seven-day moving average graphs are drawn for each topic. © 2022, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.