Construction of Interactive and Intelligible Distance Cartograms
Travel time is a critical proxy that people rely on for assessing a cost of travel from one place to another. Although information about travel time is an important part of people’s spatial information exploration, decision-making, and information analyses, seeing the exact travel time information using the current style of maps, such as equidistant or conformal maps, is not easy. This is because most of the designs use a visual channel of color to encode an ordinal level of traffic congestion. Distance Cartograms (DCs) apply a visual channel of position to encode travel time information from a single location (typically called the origin) on a map. To do so, DCs shift features on a map such that the distances on the map represent a more precise travel time between the origin and any location on the map. This visual encoding choice of DCs enables readers to visually compare travel times to places around the origin with reduced cognitive effort. Although the concept of DCs was introduced in 1960, the building of interactive DCs has remained a challenge due to the high computational cost for constructing a time space, the space that specifies the shortest travel time from an origin to the rest of the locations. In addition, previous studies indicate that the DCs’ information presentation can confuse users when significant discrepancies between the physical distances and the “travel time” distances exist. These technical and perceptual barriers have impeded researchers and practitioners in exploring the potential of DCs as a solution for everyday spatial information exploration and decision-making scenarios. In this thesis, I propose techniques that enable the implementation of interactive and intelligible DCs. To enable the building of interactive DCs, I introduce Scalable Road-network Construction (SRC) and Quadtree Time-space Partitioning (QTP), which improve the computational efficiency for a time space construction. To build intelligible DCs, I suggest Geo-contextual Anchoring Projection (GAP) and Shape-retaining GAP (S-GAP), which controls the degree of distortion. Using these techniques, I report on formative studies conducted in order to understand a series of map user interaction types that can potentially present superior spatial information exploration experience. Based on the techniques and the interaction design, I introduce three versions of a system named Traffigram, a system built to help users in experiencing Distance Cartograms. This thesis reports on the findings of a series of studies conducted in the lab and in the wild using the three versions of Traffigram. Through the lab studies, the thesis presents findings related to how DCs can improve existing techniques used for seeing travel time information (e.g., color-encoding maps) in specific ways. In the deployment study, I present findings that show initial evidence that people can use DCs in real-world scenarios and that they can identify benefits to using DCs that is sufficient to overcome the adoption barrier. The thesis concludes by discussing the limitations of suggested techniques and study findings, future work, and unexplored problem spaces that remain.