Validation page
The Validation page allows you to use a HXL schema to check a HXL-tagged dataset automatically and provide interactive feedback. The default schema is very permissive, but you can create domain-specific schemas that report errors where certain columns (e.g. #activity) are missing, blank, or out of range.
The link "Download as JSON" at the top of the page allows you to download the report as a machine-readable JSON Validation report.
The HXL Proxy also recognises a GET or POST parameter schema_content containing extra inline schema content, though there is currently no UI support for that.
Change schema: you can provide the URL of a custom schema that providers stricter checking and more-targetted feedback, and can use the choosers for HDX, Google Sheets, or Dropbox to locate it.
Minimum error level: by default, the validation page shows information and warning messages as well as errors. To unclutter the display and focus only on high-priority issues, you can set the minimum error level to "Warnings" or "Errors" instead.
The main screen (pictured above) gives an overview of the different categories of errors, warnings, and information messages found. The Proxy provides the explanation given in the HXL schema for each one, so you can customise them as desired. For each category, you can see the number of instances, and open the drill-down screen to see a full listing.
The drill-down screen lists all instances of a category of error, warning, or information message. For each one, you can see the row number, the column's hashtag, and the value that appeared in the dataset. You can select the row number to see the error in context on the View page.
If a dataset is stored somewhere that it's continuously online, such as a Google Sheet or Dropbox folder, data providers can get instant feedback on any changes by bookmarking their validation page and then refreshing it.
A customised HXL schema allows you to give very specific feedback to data providers, including recommendations on how to improve their data.