Visualization of Crowd Flows from Positioning Data
MetadataVis full innførsel
Positioning systems, both indoor and outdoor, regardless of technology used, generate large amounts of data of varying quality about where different objects are observed in space at various time instants. Such data has been proven valuable for both commercial interests and for researchers alike. Given the nature of positioning data, it is invaluable to have tools for visualizing them, with maps being a natural part of the visualizations. Various efforts have been made to generate different visualization tools for data collected at the campus of the Norwegian University of Science and Technology, but they have only made use of small subsets of the available data, partly due to how it was stored. Later, work has been done on improving the storage mechanisms, suggesting that it should be combined with the other work. This project investigates building a web application and a data processing component with the goal of providing a research platform for exploring a larger part of the data set that for example can aid in the development of data cleaning methods. The result is a set of patterns discovered using the platform and an evaluation of the efficency of using the web platform as a basis for large-scale visualization applications. Experiments show that there are certain limits to what is possible using current technologies, but that there still is a great possibility of visualizing large datasets.