The problem with those numbers, as I understood them, was they were aggregates over all time and for all locations. But, over time and locations conditions change a lot making it hard to estimate transmission rate in general, and hard to compare because the data for different variants comes from different conditions. Among the potential confounders - lockdowns and vaccines are probably the biggest. I hadn't seen your figure 12 time series data, but a time series of estimates over bounded time intervals could allow better comparison of R0 estimates I think, without needing to control for so many variables. Of course to the extent that delta "took over" due to vaccine based selection, it's hard to get a handle that way on comparative R0.

I love how in this document they present this somewhat useless data (the aggregates) and then say, but you can't use these to compare variants because these aggregates are useless. Thanks, guys! Keep up the good work! (Grrrr! Hissss!)