MACHINE LEARNING AND ANALYSIS OF TWITTER DATA: IDENTIFYING BUSINESS TRENDS AFTER COVID-19

Abstract

With the Covid-19 epidemic, there has been a great change in the routines of social and business
life. These changing routines have brought with them new needs and demands. In order for business
life to adapt to this new order and develop new strategies, current trends should be analyzed. In this
study, the most demanded business trends on Twitter after Covid-19 were analyzed by machine
learning. Textual expressions obtained through Twitter are converted into data by methods such as
natural language processing. Analyzing these data correctly makes it possible to obtain important
information that will create a roadmap about the targeted issues. Within the scope of the research, a
total of 48765 tweets with high impact were selected. Word frequency analysis was applied to the
total number of tweets belonging to the determined business trends. Within the scope of the
research, textual expressions obtained through twitter platforms were converted into data by natural
language processing method. In addition, a word analysis model based on SVM, one of the machine
learning algorithms, was used. As a result of the analysis; online food services, online sales
specialist, remote working, healthcare professionals, personal coaching, online training and
repairman have emerged as popular lines of business.