Supporting Transportation Decision-Makers with Tool Design and Data Uncertainty Visualizations

Transportation decision-makers from government agencies play an important role in addressing the traffic network conditions, which in turn, have a major impact on the well-being of citizens.The practices, challenges, and needs of this group of practitioners are less represented in the HCI literature.On the other hand, there is a gap for tools that can fulfill decision-makers' needs and help them monitor and analyze traffic conditions on the road network.Additionally, almost all transportation data have some uncertainty, and good communication of data uncertainty through visualizations can have a great effect on the quality of the decisions that transportation experts make.However, there is little knowledge about how transportation decision-makers deal with uncertain data and perceive data uncertainty visualizations.We address these gaps first through an interview study with 19 practitioners from Transports Qubec (MTQ), a government agency responsible for transportation infrastructures in Qubec, Canada.We then created design guidelines based on the results of these interviews and designed a user interface for traffic conditions analysis.The user interface design was evaluated and iterated through short user studies with 6 practitioners working at MTQ.To help transportation practitioners understand and address data uncertainty, we proposed a novel method and applied Jensen-Shannon Divergence (JSD) to quantify data uncertainty in traffic speed.We later used several visualization techniques such as color, shape, and size to create uncertainty visualizations that present the levels of data uncertainty in overlapping with the original speed data.We then conducted a user study with 11 expert and novice transportation decision-makers to explore how practitioners consider data uncertainty in their decision-making process and also understand their preferences and perceptions on our data uncertainty visualizations.In our research, we found that decision-makers who work on monitoring and analyzing traffic conditions on the road network can most benefit from research about data analysis tools and platforms that (1) provide information to support data quality awareness, (2) are interoperable with other tools in the complex workflow of the practitioners, and (3) support intuitive and customizable visual analytics.These implications can also be informative to the design of tools supporting other decision-making tasks and domains.In addition, transportation practitioners found our tool design to be useful in helping them understand and analyze their data.They also pointed out that flexible data filtering, choice of visualization selection, and data comparisons features on a single user-friendly interface can be very beneficial for viii their decision-making process.Further, our user study about data uncertainty revealed the important challenges that practitioners faced when dealing with uncertain data, such as bias in data collection, human and device errors, and lack of trust in third-party data providers.Our JSD-based uncertainty quantification method was generally well perceived by the participants to help achieve better transparency and awareness of their data.Additionally, the user study results indicated that factors such as familiarity to the representations, existing connotations of the visualization based on practice, and the goals of data usage influenced the participants' preferences on data uncertainty visualizations.In summation, the results of this thesis can be used to create better visual analysis tools to support practitioners in the transportation-decision making practice.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.