Data set

The purpose of this exercise is to develop data visualization techniques to illustrate that one variable is a mediator for a treatment effect and that this mediator effect is stronger than in other variables.

Please see attached pdf for a description of the data viz challenge for October 14, 2020.

The data set can be viewed here or downloaded as a csv file.

Goals of the data viz

The target audience for this data viz are prescribing physicians and thought leaders in dermatology. The data viz is for ‘explanatory’ purposes (not ‘exploratory’). The story is that treatment is superior to placebo in terms of patient reported outcomes, DLQI. The treatment effect is mediated by it’s effect on itch. (The reason why treatment improved DLQI so much is due to it’s improvement on itch which directly improves DLQI.) And, moreover, the other variables considered, redness and BSA, do not mediate the treatment effect on DLQI.

We want to tell this story in a compelling way that is visually appealing. Other things to consider, the clinically meaningful DLQI value is 5. (A patient who has a score greater than 10 is considered to have been impacted seriously by the disease.) A clinically relevant threshold for itch is 4. (A patient with an itch score greater than 4 is considered to be suffering greatly.) A clinically meaningful threshold for BSA is 10% and a clinically meaningful value for redness is 4. (A BSA value greater than 50% or a redness score greater then 4 is considered clinically meaningful.)

Take into account the missing data in your data viz. Let readers understand the nuances of the data as well as the message. Our objective is to be objective. We want to be transparent with the data in addition to conveying our insights into the data in a concise way.

Downloading CSV file

NOTE to download a single data set as a csv file, click on the raw button and save the file. The following link describes the process in further detail.