Exploring the reach of departments of transportation tweets: What drives public engagement?

State Departments of Transportation (DOT) utilize Twitter frequently in order to disseminate critical information regarding traffic conditions (e.g., accidents, roadway closures, and congestion) to public. Even though factors driving the effectiveness and reach of private social media accounts have been studied extensively in the literature, they are still not fully understood by federal and state transportation agencies in the context of information dissemination using Twitter. In this paper, Twitter interaction analytics such as engagement rate and impressions, and other exogenous and endogenous variables are utilized to identify those factors that influence the effectiveness of this information dissemination. For this purpose, a methodological framework is proposed involving: (a) machine learning and naïve Bayesian techniques to classify and extract information from the DOT tweets, and (b) a selection model to identify the significant endogenous (e.g. time-to-post, tweet click, tweet engagement) and exogenous (e.g. demographical, socioeconomic, and land use characteristics) factors that drive the engagement rate as an indicator of information dissemination effectiveness. This framework is implemented via a case study application based on the Twitter account of Florida Department of Transportation (FDOT) District 3 region. Results reveal that factors such as tweets’ time-to-post, tweet analytics, roadway accident information, and demographic variables drive engagement to tweets. As such, this study can assist transportation agencies to calibrate their plans and policies towards providing better and faster information to the public, and improving the effectiveness and reach of their social media accounts.