The Accessibility team are using a number of automated services to check pages for their accessibility. They have made an assessment of ten of these tools here.

What they want to know is:

• What are the best combinations of the different tools
• How does this change when you consider different scoring criteria?

The three scoring criteria are:

• error found
• error found + warning reported
• error found + warning reported + manual inspection

• Hash the various accessibility fails to make them easier to work with
• Create three dataframes and binarise for the three success criteria
• Compute rowsums for all the possible combinations of accessibility tools
• Visualise this data somehow

• Test 16 and 81: 'Text language changed without required change in direction' were replicated in accessibility_fail_data.csv. The test at line 81 was appended with a 1 in teh csv to solve this issue.