Authors: Caitlyn McColeman, Mark Blair
Abstract: Multivariate data visualization is as challenging as it is necessary: most real-world data problems involve complex, interrelated variables. This paper explores the efficacy of candlestick plots by recording participants’ forecasting performance after viewing hundreds of such graphs. In addition to the standard candlestick chart, we test an adapted version of the chart, where the data are represented more intuitively (i.e. an up arrow means “the stock closed higher”) than in the standard chart (i.e. a green box means “the stock closed higher”). Performance on the adapted candlestick chart is better: participants were more accurate in their forecasting predictions when the data were represented more intuitively. This suggests that our innovation in multivariate financial data visualization helps people make better decisions.