Digital News Transformation on Education in the Most Affected Country by COVID-19 Using the Topic Modeling and Sentiment Analysis

This study focuses on examining US newspaper articles regarding education from the two major news channels the New York Times (NYT) (N=29.682) and Washington Post (WST) (N=44.308) in period 1 January, 2020 - 19 March, 2021, and splitting them into three stages. We employed Latent Dirichlet Allocation topic modeling and sentiment analysis to depict the overall picture of the data set. Our method flow chart included start, data preparing, data analysis, and result. We used Python to call Google API to calculate the sentiment analysis score. There is a difference in the frequency of the occurrences of the education theme in NYT and WST in the three stages, where NYT relatively dominates. Keywords related to education that appear on the NYT and WST include school, child, parent, student, child, family, feel, and home. Sentiment analysis scores on all themes in NYT and WST were generally in the neutral categories, while the direction from stage one to stage three tends to be more positive. This study could be useful to assist education policy makers in determining the right decisions in the implementation of quality education after the COVID-19 Pandemic era.