TrendTracker: Temporal, network-based exploration of long-term Twitter trends

TrendTracker is a web application for the network-based and temporal exploration of long-term social media trends. Topical trends, represented as a series of hashtag co-occurrence networks, can interactively be explored while the user is provided with detailed trend analysis insights. This approach has several benefits compared to alternative trend visualization and exploration methods, such as ranked lists of trending keywords, as it provides the user with additional context-sensitive information. To showcase the TrendTracker application, we leverage a Twitter dataset of German political actors and demonstrate the system's capabilities in various ways. For example, the user is able to investigate a single trend from multiple perspectives, such as the trend's temporal development over time, including its topical shifts and changes in popularity. Also, given the network-based trend visualization, the user can intuitively understand the different facets of a trend and how these are interrelated. Thereby, individual hashtags and relationships can be tracked over time as well. Furthermore, the TrendTracker application allows the user to compare trends. This way, differences in the trends' temporal evolution or topical alignment can be uncovered. The demo is publicly available via the following URL: https://trend-tracker.ifi.uni- heidelberg.de