Be careful with your data transformations!! The combination of your variant filter with the expand offset is distorting the data. Expand normalizes the data for every step in the x-axis (for every date) to 0 - 100%: in January 2021 the Alpha variant represented 30.77 percent, while Iota represented 15.38 percent of the cases, while in your chart it looks like it was 65 and 35 respectively.

See my comment on the offset: "expand" above. The labels on each area are a great addition and improve readability. BTW, Plot has recently added a new type of mark, for legends: color: { legend: true } Although I like your labeling better. The waves' height is a bit too exaggerated, filling the cell's width can give you a better aspect ratio: width: width