Sridhar Venkatraman, a subscriber of the e-paper from Milton Keynes in the U.K., wrote to us with a few interesting questions about data visualisation in news stories and on the oped page. He said: “First, hats off to the graphics team for making interesting charts and maps and introducing variety to the readers. The charts and maps help in understanding numbers better than tables do. However, the circle-style graph is becoming repetitive. And in some cases, the circles are too small and the text on them is so tiny that I have to zoom in to read. If I zoom in, the context of the entire graphic is lost and I can see only one or two points in close up. As a reader, I am not interested in zooming in and seeing every single State in the country.” Mr. Venkatraman then listed seven data stories to substantiate his argument.
Using scatter plots
The reader makes an important point about the inherent problems with data visualisation. The charts which he refers to are called scatter plots. The data team uses scatter plots specifically for those stories that require multiple variables to explain the complex nature of interplay.
The data team explained the rationale for relying on scatter plots rather than the usual pie charts for the stories cited by Mr. Venkatraman. For instance, in the story “Karnataka sees 300% jump in FDI inflows” (July 24), the data team used a scatter plot to show changes in FDI equity inflows between 2016-17 and 2017-18, respectively, for major States (represented by their Reserve Bank of India circle offices). The scatter plot sought to answer some questions: Which were the States with the highest and lowest FDI inflows? And among States within the high and low FDI inflow categories, which were the ones that registered an increase or decrease and what was the magnitude of those changes?
“By plotting the variables on a scatter plot and setting the graph with a trend line, we were able to isolate States above the trend line (increase in FDI) and below (decrease in FDI). With varying circle sizes in the plot (called bubbles), we indicated the percentage increase in FDI inflow. And by plotting the States on an x-y axis, we managed to show which States had the highest and lowest FDI inflows. All these were conveyed in a 2.5 column and 5 cm space. With the text, we managed to restrict the graphic to a rectangle of 3 column and 10 cm space,” said a data team member.
The data team also explained a data story, “Mutually dependent”, that appeared on the oped page in the ‘Data Point’ section on July 24. “We used a scatter plot to show that a higher number of patents are granted in States that have a flourishing start-up environment and vice versa. Data points have a fixed space of about 2 column x 9 cm. If we had not used scatter plots, we would not have conveyed the entire picture. The Data Point would have only managed to show either the number of patents granted per State or the number of start-ups there. It would have been a simple visualisation — a map of States or a bar chart that showed the States with the highest and lowest number. But it would have missed the bigger picture. The scatter plot, therefore, was again chosen specifically to enhance the information,” he said.
Content and visual appeal
Space is a major constraint in print stories as the newspaper has to accommodate many stories on one page. The scatter plot is a useful device to conserve space and maximise information without missing out any important component. According to textbooks, data visualisation is defined as “the use of computer supported, interactive, visual representations of data to amplify cognition”. While the primary focus of these academic exercises is to find out how to make numbers accessible, the aesthetic experience of the reader and the ease of reading is not given due consideration.
It is true that content comes first in a newspaper and visual aesthetics next. However, in the case of infographics, it is the visual appeal that invites the readers to a data set. Increasing the space for data stories may be an answer to this conundrum.