Visualizing Temporal Data
The following graphs show three different ways to depict temporal data: a line graph, a connected scatterplot, and a Ganntt bar graph. A line graph, with a timeline on the x-axis, allows the audience to quickly detect temporal patterns. However, when two phenomena are shown in the same graph, the audience might get confused that those phenomena have the same unit. Thus, chart makers have to emphasize the right-side x-axis to make sure audience understand that those phenomena have different units. The connected scatterplot is the one in which disadvatanges might outwieght advantages. Ususally a scatterplot is to allow the audience to quickly tell whether there is a positive or negative relationship, but connect the dots to emphasize the time dimension might even obscure the ability of the audience to see the relationship. Indeed, from the line graph, the audience can see both the increasing trend and the positive correlation between pitchers and stikeout, whereas it is hard to detect from the connected scatterplot of both trend and relationship. The Ganntt bar is good at depicting durations of different observations within the same phenomenon. However, in ganntt bar, time is merely a unit and not a phenomenon and thus it is not useful if the purpose is to show temporal pattern such as cycle and seasonality.