Evaluating the traffic condition in Jakarta during COVID-19 pandemic: Social media text mining.

During the COVID-19 pandemic, the local authority in each city was implementing different approaches to limit the virus spread within the population. Jakarta, as the city with the most reported cases in Indonesia, reinforced the lockdown policy to limit the activities of its people. This regulation has been reported by news outlets as the main factor which improve the traffic condition in Jakarta during this period. The present study proposed a text mining approach to explore this phenomenon in the social media realm, especially Twitter. The tweets used in this study were gathered from the Jakarta police department's official Twitter account, @TMCPoldaMetro. This account regularly reports the traffic condition from several locations across Jakarta. The proposed method extracted the location and its traffic condition reported on the tweets. Then, the weighted scores, from zero to two, were calculated to quantify this traffic condition during the lockdown period (10^th of April - 31^st of December 2020). These scores are compared to the same period from the two previous years, as well as the previous period within the same year (1^st January - 9^th of April 2020). Firstly, a series of t-test was conducted to statistically measures the traffic condition differences in general. The result of this study indicates a significant traffic improvement during the lockdown period in overall reported locations. Furthermore, chi-squared tests were also performed to measure traffic improvement in each specific location. Surprisingly, several locations which considered as office district was reported no traffic improvement. [ABSTRACT FROM AUTHOR]