The HTML report has proven a valuable asset for troubleshooting on both the lab and analysis side. I think there are two ways the report could be improved, though.
Currently, FastQC is run on each sample's R1 and R2 separately. We should look into performing a single MultiQC run on all of the samples together and linking to a single MultiQC report.
Currently, read length distributions are plotted sample-by-sample before and after merging. This is displayed as a grid of histograms, one per sample (or two for R1 and R2 pre-merging). We should look into replacing these histograms with a single ridge plot (code examples here).
The HTML report has proven a valuable asset for troubleshooting on both the lab and analysis side. I think there are two ways the report could be improved, though.