00000ctmaa22000001a 4500 UP-1685675941123958482 Buklod 20190724152945.0 a r |||| u| ta 190724s xx d r |||| u| (iLib)UPTAC-00019221837 UPVTC eng LG 993.5 2019 C66 O64 Opetina, Patricia Nicole Evaluation of visualization approaches for analysis of event logs for post-election audit. Tacloban City Patricia Nicole Opetina 2019 ix, 73 leaves illustrations. Undergraduate thesis (B.S. Computer Science) -- University of the Philippines, Tacloban. With the use of vote-counting machines (VCM) in Philippine elections, tracking of VCM operations by means of event logging is crucial for transparency, maintenance, and management of these machines. While VCM logs contain a vast amount of data, it is not obvious at once what kind of meaningful information can be digested from it. In extracting insights from these logs, presentation and modeling play a vital role. This work explores the potential of data visualization as an aid in answering election-related questions based on data extracted from VCM logs. Different visualization techniques such as table charts, line graphs and circle segments were applied on a dataset containing artificial VCM logs that were generated based on VCM behaviors observed from previous Philippine elections. To assess the performance of these techniques, a user-centered evaluation is employed, by means of a usability study. The usability study involves the participants performing a set of tasks by using the aforementioned visualization approaches to achieve goals such as VCM failure detection and events pattern recognition. Preliminary results showed that the effectiveness of each visualization type is task-dependent. That is, some tasks were performed better by using certain visualization types while other techniques did not do well on that same set of tasks. Refinements to the prototypes were made to improve their overall performance in the final usability study. The results of the final usability test showed that among all the visualization techniques experimented, the enhanced line graph and colored table chart are the most suitable prototypes in visualizing VCM event logs because they characterize the identified attributes needed for successful VCM log visualization. Ultra, John D. adviser. FI UP UPTAC UPTAC LG 993.5 2019 C66 O64 Thesis