Forum Replies Created
-
Posts
-
I have mostly encountered box and whisker plots in research articles, not often in dashboards. I don’t think they are easy to understand and to communicate data to a public audience. It requires the audience to have an understanding of several statistical terms and know what their interpretation are. I think they usually require more interpretation than the concise data summaries that are usually communicated in dashboards.
However, I frequently encounter histograms. Epi curve, the graph of disease onset during an outbreak, is usually a histogram. They can work well for communicating counts over time.
I agree that a horizontal bar chart would be a good way to display a ranked variable. I think the reader would easily understand the top to bottom ranking system and it would mimic score charts and other ranked tables and charts that we see. However, sorting the bars according to the values would be key to this graph. If the service categories appeared in alphabetical or some other order, then it wouldn’t be as clear to the reader.
Both zones would have the same service category as their most frequent claim type, curative care visits. The Northern zone has more (1110) than the Eastern zone (987), but they are the most frequent claims for both areas. The least frequent service category in the Northern zone is speech therapy (1), while in the Eastern zone it is both inpatient obstetrics (2) and inpatient medical (2).
I agree that a line graph would be a good choice for this data, but I also think that a bar chart/histogram would be a good choice as well. It would show the change in membership volume well from month to month. Additionally, the Flores del Mundo data are not continuous data, they are cross-sectional, point-in-time assessments of the membership once per month. Because we are measuring membership at regular intervals, a bar chart could be a useful visualization of these data.
However, once we add the stratification by member type and region, it could be difficult to compare a bar for each group during every month. We could try a stacked bar chart, since we are only comparing two or three categories, but if a group is small, it could be difficult to see. This may be the point where I would switch to a line graph so that the reader can easily see the different groups on the same graph.
Using Few’s definition of a dashboard, “a visual display of the most important information needed to achieve one or more objectives; consolidated and arranged on a single screen so the information can be monitored at a glance.” I think that the WHO maternity dashboard is a dashboard. It’s consolidated on a single screen, only specific metrics are included, the data is limited to only the targets and indicators over time, and the color coding helps provide that “at a glance” interpretation. Further, according to the article, this dashboard is for use by local healthcare providers, so there isn’t a need for definitions, simpler text, or additional explanation of the the indicators, since they are likely already familiar with them. I agree with previous comments that the dashboard could use some design assistance, but overall I think it meets it’s goal of being a dashboard.
