Visualise data with different sample sizes

Question

I did a literature review and ended up with a lot of data in the kind of "10 out of 20 studies found x while 4 out of 8 studies found y".

To be more precise; I looked up the contribution of seafood to environmental impact categories in different studies. Here is a short extraction of my data:

I would like to represent the data somehow now. For example I would like to have a graph that shows how many cases of "Fishing" contributes "33%" compared to "Processing", "Packaging" and so on.

How can I deal with the different sample sizes? Should I select a minimum number of studies to include and if yes, is there any criteria to set that limit?

I don't need to do statistics but would like to have some simple graphs to allow readers a very quick first overview.

Answer 1

A modified bar chart with bar width (height in this case) representing the number of articles may work. This one is rotated by 90 degrees to accommodate the long axis labels.

There are many other ways as well:

1. Pictogram (use a document icon for each literature) or a barchart that is formatted like a population pyramid, one side is present, the other side absent.
2. Combine bubble chart and pie chart by using the bubble size to represent number of literature and the slices to represent present/absent evidence.
3. As a modification of 2, you may also consider squarified tree map.