Just some constructive feedback:
The graphs are in my opinion far too bloated. You should have really cut down the amount of content on the graphs. Anything above a 580/1660 SHOULD NOT have been displayed. It's worse for mobile users and makes it harder to read. To further clean up the cards not focused in the review but are being displayed, only 1080p performance should be shown. We don't need to know a 1660's 4k or 1440p performance to make a conclusion on 570 vs 1650.
Also the review is focused on the 570 vs the 1650 cards and the colors(blue/green) only show for the 570. Now it's still not difficult to distinguish between the two, but two better options could have improved the graphs overall. 1) If ignoring the above feedback, display only 570 vs 1650 and use red/green colors. 2) Keeping how it is with more cards displayed, use the red/green colors to distinguish them for both 570/1650.
You did a great job with the Division 2 graph for example. That is almost exactly how they should be. Removing the 1440 and 4k results would make it exponentially better.
Lastly personally the graphs I think need a modern overhaul. They are aesthetically unmodern and older looking. Elevating the look of your graphs will really help to make the review shine.
Also a sidenote: You really should only benchmark Total War Warhammer 1/2 in DX11 mode. DX12 has issues and performance wise outside of benchmarks is unstable and inconsistent. There have been hundreds of threads on their personal forums with people complaining for years and switching to dx11 has solved basically everything. It is not a reliable source of results and misinforms the reader.
Other than that, as usual, the actual information learned from your articles is still as valuable as ever. The 570 still has some grunt left and compared to the 1650 its by far without question what everyone in that price range should be buying.