Adaptive Visualization of Tourists' Preferred Spots and Streets Using Trajectory Articulation

Walking tourism, in which regional resources are organized with interesting themes, can provide visitors with original local walking experiences. Our project aims to collect user data through a mobile application and explore potential geographic resources such as appealing spots and streets for improving city-scale tourism. A density map with GPS trajectory data is one of the easiest ways of visualizing them without any modeling costs. However, both user and technical factors make it difficult to interpret the heatmap in a detailed and concise way. Specifically, analysts have difficulty in deciphering the areas of real interest based on the heat map using the data as areas associated with high density of GPS locations may not be solely due to their attractiveness, e.g., rest areas. In addition, the heat map that does not retain the topography of the streets cannot achieve hot street visualization. In our research, built-in smartphone sensors are employed to distinguish multiple user contexts (e.g., stopping / walking and indoors / outdoors) during their walking tours, which equalize the degree of inherent density biases in each GPS trajectory and add attributes to each location point. Our analysis software accumulates the processed trajectories and generates a density map by applying different weight rules (e.g., a street-oriented rule and an indoor-oriented rule) based on semantic attributes and analytical requests. Our mobile cooperative approach realizes adaptive heatmap generation to the analyzer's expectations, that is, concise hot spots visualization and hot streets visualization.