New York City Street Cleanliness: Applying Text Mining Techniques to Social Media Information

Abstract

This study examines social media information within the context of social network theory by utilizing text mining techniques and machine learning algorithms to analyze the tweets and use the information as a proxy to measure the performance of the municipal services. The objective of this study is to provide the government an alternative measure by examining user-generated content, identifying temporal trends and patterns of street cleanliness, evaluating whether people’s opinions are consistent with NYC cleanliness ratings, and assessing the performance of municipal services via sentiment analysis. The results indicate that the analytical approach presented in this study can systematically analyze social media information, and the information can be used as an alternative measure or supplementary evaluation tool to validate the government’s cleanliness ratings and evaluate the performance of the municipal services. This study extends the research on accessing social media feeds to gain valuable insights regarding municipal services’ performance. The analytical approach can be generalized to many other areas where social media information can be utilized.