The general idea here is the we can use the data set hoster as a tool to literally walk the datasets that are in the data set hoster. For the similar-artist query page it should be trivial to repeat the similar artist query from one of the results of that page.

To make this possible, when the dataset hoster shows a row of results, for example a recording_mbid, right next the the MBID should be a dropdown. This dropdown allows the user to use this MBID as the input for a new query on the dataset hoster, ideally in a  new browser tab.

For example on our artist similarity page ( https://labs.api.listenbrainz.org/similar-artists?artist_mbid=8f6bd1e4-fbe1-4f50-aa9b-94c450ec0f11%2C1c70a3fc-fa3c-4be1-8b55-c3192db8a884&algorithm=session_based_days_7500_session_300_contribution_3_threshold_10_limit_100_filter_True_skip_30 ) we have two sections of results. The artist summary and their similar artists.

To the right of each of the artist_mbids, the dropdown should have one item for each type of query on the dataset hoster that accepts artist_mbid as an argument. If the user selects that item a new browser tab opens and that query is run. If there is another argument like  algorithm which was also present in the current query use the same value to start the new query.

There will be cases where the result page may not have all the data items needed to carry out the next query. In this case, a new tab should be opened and the known values should be filled out, but the query should not be attempted to run, since it will fail.