there are slightly different results for "mean" and "sum", aggregation, i think "sum" might be more appropriate for "most revenue". but the chart is good!

i think you would benefit from correctly aggregating the orders :) the solution is shown in this video https://www.youtube.com/watch?v=wFLx3vQdstw at 4:20 mark

also a challenge for you: can we get human-readable week day labels?

oh, you've aggregated the state! you should've aggregated orders! instead (Y)

for human readable days of week here is the solution outline: it's only possible to do in code add a configuration for facet x encoding, similar to how you now have "color: { legend: true }" inside facet x config add domain and range, two arrays your domain would be the 0,1,2 etc — data values your range would be days of week like "Sun", "Mon" i hope the outline brings you closer! :)

1. good that the chart is showing all the orders in one view 2. bad that it's flat, which means all information has collapsed into the overplotted mess 3. bad that the color palette has so many categories it's not useful proposed solution: split names into type and size (🧿 4H tutorial should give an idea how to do that with Array.map(), struggle for a while and if still stuck i can help you do that), then use facet-x for sizes and facet-y for types this will untangle most of the overplotting if not all of it