Designing visualizations to identify and assess correlations and trends: An experimental study based on price developments

Lisa Perkhofer, Conny Walchshofer, Peter Hofer

Research output: Chapter in Book/Report/Conference proceedingsConference contributionpeer-review


Alongside the increase in available data, long histories, and the need to look at unconventional investment strategies (high risk and low risk by focusing on parallel or opposing stock price developments), multiple visualization options have emerged. This is caused by the ability of visualizations to provide insights such as an accurate and efficient assessment of possible correlations and trends. This study focuses on an optimal way to visualize correlations between two officially listed price developments (stock prices, indices, and commodity goods). In this regard, the choice and the design of the visualization used can influence decision accuracy substantially, however, explicit effects on visualization use and design choices are mostly lacking. To fill this gap, this study tests two highly recommended visualization types (a scatterplot and a parallel coordinates plot) and three concrete design features (regression line – yes vs. no; color – mono vs. multi; interaction – filter vs. select). Although the results indicate that scatterplots outperform parallel coordinates plots in all design conditions, parallel coordinates plots are less affected by deviations from a normal distribution (measured by kurtosis and skewness) and with increasing experience they might be equally effective.
Original languageEnglish
Title of host publicationProceedings of the 17th Finance, Risk and Accounting Perspectives Conference (FRAP)
Publication statusPublished - 2019
Event17th Finance, Risk and Accounting Perspectives Conference (FRAP) - Helsinki, Finland
Duration: 23 Sept 201925 Sept 2019


Conference17th Finance, Risk and Accounting Perspectives Conference (FRAP)
Internet address


  • Big data visualization
  • visual analytics
  • scatterplot
  • parallel coordinates plot
  • correlation identification

Cite this