COLORED SLOPE GRAPH: VISUALIZATION OF TIME SERIES DATA
thesisposted on 05.05.2020, 13:20 by Apurva Nagarajan
Time-series data draw extensive attention from many research domains, such as financial and biomedical engineering. Researchers often face the difficultly of visualizing multiple time-series data simultaneously. The predominant techniques suffer from visual clutter either due to over-plotting or an overwhelming number of small graphs that carries a high cognitive load. This research study proposes a Colored slope, a combination of Tufte's slope graph and time-series heatmap, to visualize multiple time-series data at once, balancing scalability and accuracy. Colored slope inherits the complementary advantages from each method, regarding visualizing temporal changes within a period and identifying precise values. The efficacy, effectiveness, and graphical perception of the Colored slope on visualizing multiple time-series data with publicly accessible stock data were evaluated and compared it to popular time-series visualizations, including line graphs, time-series heatmap, and small multiple spark area graphs. Analyzing the experimental data, this study concludes that Colored slope contributes to (1) allowing users to identify the extreme values (maximum and minimum), co-variability, the general trend of the data, and rate of change effectively for an optimally large number of visual marks(time-series data); (2) capable of display more instances of time-series data with a less visual clutter problem. Finally, several possible applications and limitations with the Colored slope are demonstrated.