Click the Original, Code and Reconstruction tabs to read about the issues and how they were fixed.

Original


Source: Nielsen Audio Today (2019).


Objective

The original data visualization is for a marketing report on audio preferences for the American audience. The original graph’s purpose is to show which type of platform can reach more Americans weekly. In addition, it intends to provide insights on choosing the right media type to reach the right audiences and recommending audio-based advertising solutions to potential advertisers for the broadest audiences (Nielsen 2019).

The visualisation chosen had the following three main issues:

1. The original graph was unable to highlight the finding.

Position of data can let viewers know the difference and emphasize the contrast (S. Mittelstädt, A. Stoffel and D. A. Keim 2014). This graph is supposed to show radio can reach relatively more Americans than other platforms. However, it just grouped all populations by one platform rather than provide a strict forward comparison between radio and all platforms. To improve on that, all the data should be grouped by population and compare which platform can reach more people for each age group.

2. The original colours in shades were unable to highlight the key patterns.

These four populations used darker to lighter green colours to show their connectivity. In an ideal data visualization, colour should help viewers to identify the difference between data groups (Stephanie E & Ann K. E 2016). However, one of the populations, “ADULTS 18+” represents the average of another three population groups; it has no continuous relationship to others. Thereby, it is better to separate this category to make it more understandable.

3. The original graph has no clear axis label.

For a better interpretation, a graph with one x-axis and one y-axis can benefit the viewer in understanding the data (Stephanie E & Ann K. E 2016). This graph shows six media types on the left-hand side, but no axis label describing their purpose. The x-axis label “Weekly U.S. Reach (Percent of Population)” is placed on the top, between the x-axis and y-axis, which can create confusion about understanding the graph. Therefore, putting a y-axis label next to the media types is suggested and relocating the x-axis label slightly more to the right, near the percentages.

Reference

Nielsen (2019) Audio Today 2019, accessed 17 July 2022. https://www.nielsen.com/insights/2019/audio-today-2019/

S. Mittelstädt, A. Stoffel and D. A. Keim (2014) Methods for Compensating Contrast Effects in Information Visualization: Methods for Compensating Contrast Effects, accessed 19 July 2022. https://rmit.primo.exlibrisgroup.com/discovery/fulldisplay?docid=cdi_crossref_primary_10_1111_cgf_12379&context=PC&vid=61RMIT_INST:RMITU&lang=en&search_scope=EverythingNOTresearch&adaptor=Primo%20Central&tab=AllNOTresearch&query=any,contains,Methods%20for%20Compensating%20Contrast%20Effects%20in%20Information%20Visualization&facet=citing,exact,cdi_FETCH-crossref_primary_10_1111_cgf_123793&offset=0

Stephanie E & Ann K. E (2016) Data Visualization Checklist, accessed 19 July 2022. https://depictdatastudio.com/checklist/

Code

The following code was used to fix the issues identified in the original.

library(ggplot2)
library(tidyr)
data <- data.frame(PLATFORMS = c("RADIO", "TV","SMARTPHONE", "PC","TV_CONNECTED_DEVICE","TABLET"),
                   ADULTS_18_PLUS = c(92,87,81,54,52,46),
                   P18_TO_34 = c(90,75,87,55,60,38),
                   P35_TO_49 = c(94,89,89,59,62,51),
                   P50_PLUS = c(91,93,71,50,40,48))

data$PLATFORMS <- factor(data$PLATFORMS, levels= data$PLATFORMS[order(data$ADULTS_18_PLUS)])
com <- gather(data, key="measure", value="value", c("ADULTS_18_PLUS", "P18_TO_34","P35_TO_49","P50_PLUS"))
p1 <- ggplot(com, aes(y=PLATFORMS, x=value,fill= PLATFORMS))+
  geom_bar(stat='identity')+
  facet_wrap(~measure, ncol=1)+
  ylab("PLATFORMS")+
  xlab("WEEKLY U.S. REACH (Percent of population %)")+
  coord_cartesian(xlim = c(30, 100))+
  ggtitle("COMPARING WEEKLY REACH")+
  scale_fill_manual(values=c("#8e44ad",
                             "#337ab7",
                             "#5cb85c",
                             "#5bc0de",
                             "#f0ad4e",
                             "#d9534f"))+
  theme(axis.text.y = element_text(size = 7),
        axis.text.x = element_text(size = 7),
        plot.title = element_text(size = 15),
        axis.title.x = element_text(size = 10),
        legend.text = element_text(size = 8))

Data Reference

Nielsen (2018) Total Audience Report Q4 2018, TV = Live + DVR/Timeshifted TV TV Connected Devices = DVD, Game console, Multimedia Device, VCR, accessed 17 July 2022. https://www.nielsen.com/insights/2018/total-audience-report-2018/

Reconstruction

The following plot fixes the main issues in the original.