#100things Create a dashboard
Dr. Alexander Schacht
Author, Speaker, Podcaster, Leadership Trainer. Fear is a reaction. Courage is a decision. The Effective Statistician! Medical affairs/RWE/HTA expert statistician.
Depending on the industry you work in or the experience you gained, creating a dashboard might be something that you have never done before as you mostly work on static reports.
We can interpret the term dashboard in many ways, but let’s think here about an interactive exploratory visualization. Even when the data is static, interactivity may help a lot with understanding the data and with easier delivery of the “report”. In many situations, we display the same variables in a certain way multiple times by combinations of other variables.
Think for example about a scatterplot where you can restrict the data displayed by various other variables. Instead of delivering each combination of scatterplots, it’s much more efficient to create an interactive version.
If you have not worked on this before, you’ll understand:
- Where to put the parts of the visualization to select the subgroups
- How to make them obvious to use
- How to adjust axis etc to deal with different ranges of variables
- Why a “reset” button may be helpful
- How to roll-out an interactive graphic in your organization
- How to get feedback on an interactive visualization as opposed to a static one.
Once you have made yourself familiar with simple dashboards with just one display, you can learn about the creation of more complex dashboards including multiple displays to understand different related parts of the data at the same time.
In either way, it’s good practice to start with the layout of the dashboard using a rather low-tech approach like post-it notes. When each post-it note represents a unique element of the dashboard, you can more easily rearrange, discard or add these as opposed to starting directly programming them. The ability to iterate different designs quick will help you select the best one among many alternatives. It also avoids the bias of sunken cost. For this reason, we tend to stick to a solution, in which we have invested a lot of time in. Thus, we’ll more likely discard easily generated poor options than those who took us a lot of time.