Visualising Multi-Strain Data

Question

[rowe]

Hi everybody,

I have data of ten strains, which are filtered with different methods (one following the other). The goal of the visualisation is to show the amounts of filtering per method in one figure. The first thing that comes to mind is a 10-level donut chart (I don't know how yet). Are there any better options than this?

The workflow:

Total fastq  ->  method 1 ->   method 2  ->  method 3

The data structure:

Strain    Total    Method 1   Method 2    Method 3

RHS       100%     80 %       6%              12%
EL        100%     73%        7%              20%

So basically two questions:

• How can I make a 10-level donut chart in R?
• Are there better options? How?

Answer 1

[B. Arman Aksoy]

For question 2:

I wouldn't do a 10-level donut chart; because the inner rims will be really hard to distinguish and quantify for the eye, especially if you want to compare methods (hence different donuts with each other). To better decide on the visualization method, I think you have to decide on what message you will be conveying to the reader.

It seems to me that, a grouped bar-chart where you group your statistics by the method will be sufficient and easier to read. This will, for example, help check what is the general effect of applying that method on any strain. If you want to emphasize, for example, some strains are more filterable than other, then I would think about grouping the bar charts by strains instead.

If your filters are mutually exclusive, then how about a stacked bar chart?