Real-Time Traffic Event Detection From Social Media

Abstract

Smart communities are composed of groups, organizations, and individuals who share information and make use of that shared information for better decision making. Shared information can come from many sources, particularly, but not exclusively, from sensors and social media. Social media has become an important source of near-instantaneous user-generated information that can be shared and analyzed to support better decision making. One domain where social media data can add value is transportation and traffic management. This article looks at the exploitation of Twitter data in the traffic reporting domain. A key challenge is how to identify relevant information from a huge amount of user-generated data and then analyze the relevant data for automatic geocoded incident detection. The article proposes an instant traffic alert and warning system based on a novel latent Dirichlet allocation (LDA) approach (“tweet-LDA”). The system is evaluated and shown to perform better than related approaches.

Exploring the Potential of Twitter to Understand Traffic Events and Their Locations in Greater Mumbai, India

“…Tweets collected using a spatial extent are geotagged whereas tweets collected by keywords or by following some specific user accounts are typically not geotagged. Some researchers [8], [22] used multiple strategies to collect traffic related tweets. For example, Wang and colleagues [8] collected tweets from official accounts, using pre-defined road names and using circular search areas along the road network to collect geotagged tweets near roads.…”

“…However, they lack any semantics describing the nature of events, which are thus algorithmically inferred. A further potential source are actively generated data, the subject of this paper, where individuals actively report on events through social media posts or microblogs [63], thus acting as social sensors, providing information about ongoing events, ranging from cultural festivals [7] through natural disasters [6] to the subject of this paper, traffic events [8], in a dynamic way [49]. The value of these data lies in their semantic richness: a single tweet reporting an accident at a specific location provides rich information about current and historical traffic conditions.…”

“…There are three primary sources of location information in tweets, the user's home location mentioned in their profile, location provided in the tweet metadata, and location mentioned in the tweet content [26], [51]. Much previous work has used the location information from tweet metadata, in terms of coordinates, e.g., latitude, longitude [8], [15], [17], [50], although only a very small proportion of tweets are actually furnished with this information (Figures ranging from 0.1% to 3% of the total volume of tweets have been reported [24], [72], [73]) and Twitter announced in June 2019 that it would no longer share precise locations in metadata. Since profile-based information is essentially static, extracting loca- 1 https://twitter.com/ This work is licensed under a Creative Commons Attribution 4.0 License.…”

Framework for fusing traffic information from social and physical transportation data

“…In previous studies, some researchers used joint words, such as “from…to” and “between…and,” to identify such location information [54,57]. However, this rule-based method lacks flexibility, and the locations may not be the exact locations, but rather the nearest landmarks that social media users can refer to.…”

“…Wang et al . [32] proposed a traffic alert and warning system using the latent Dirichlet allocation method to identify transport-related social media data. Gu et al .…”