User-centric Web-based System for Visualization of NIS-data for Layman Users

Abstract: Spatial data is playing a bigger role within many fields outside of Geographic Information Systems (GIS) and spatial analysis. With more and more users with varying levels of previous spatial analysis experience using this kind of data, there is a growing demand on how this information is best presented to the user. This user-centered design is an increasingly common theme in other adjacent fields, but is still in its infancy in the field of GIS. Currently there is no obvious generalized solution that provides the answer to how to present data, no matter if it is spatial or not. How to present data in a smart and comprehensive way is still an everyday challenge across many fields. The objective of this thesis is to create a prototype of a web based Network Information System (NIS) where the layman user is in the center of the entire design process. This includes both the actual visualizations, but also the choice of tools and the interface design. The prototype is designed around the role of the customer service representative in a utility company using a NIS. This type of layman user is the kind of user that today works in a system that is designed with neither their role nor their GIT experience or training in mind. From this prototype, the efficacy of different visualization techniques on layman users is evaluated, producing more general guidelines for user-centered development directed at layman users. The first step of this user-centered design process is to understand the user. By interviewing users of the system, their current work flows and opinions of their current system are better understood. From this, information about which tools they need, which current features work well and which need revising is gathered. Based on this, a mock-up is created which is then transformed into a prototype. Finally, the prototype is evaluated by the target audience with comments on a presentation as well as a larger survey. The results show that in general the prototype is well-received with regards to existing functionality and how it is presented through the interface design. The implemented visualizations are well understood by most of the expert users, but are less successful with the layman users in the survey. Especially with regards to the icon choices and other point representations, there is a discrepancy between the intended visualization and the perception of the survey takers, which may partly be due to the lack of context presented. An appreciated fact is that the functionality implemented in the prototype is tailored to the requirements put forth by the users. User-centric design processes depend heavily on the developer's understanding of the user and their needs. This is as true for functionality as for visualizations, where familiarity and associations can be both beneficial and detrimental, depending on how well understood they are. Using icons to represent objects is very efficient, as long as the context and the meaning of the icons themselves are well defined and conveyed. Finally, it is imperative to not throw too much information at the user. Whether in the shape of too many tools and options, or by displaying too much on the map, the same clutter-problem occurs. When presenting a situation or a scenario, the core message cannot be obfuscated by unnecessary features, functions or choices.