Three things to add...
First, there are spreadsheets and tools for dialing in the MAF scaling in both open-loop and closed-loop modes. I prefer to force the ECU to run open loop (details here
). That you can use one method to dial in the whole MAF scaling from top to bottm.
Second, that one method is just a fairly simple Excel spreadsheet. Given a data containing your real AFR (from a wideband sensor) and the ECU's target AFR (from the "final fueling base" parameter), make an "AFR Error" column (= real_AFR / target_AFR). Then make scatter plots of error over MAFv and error over RPM. Everything I know about Excel I learned from trying to visualize the information in my data logs. Learning Excel was a fringe benefit of learning tuning.
Third, generally you only need to look at error over MAFv, but occasionally you'll find places where the same MAFv corresponds to both positive and negative errors, depending on RPM. You can fix that by fudging the fuel table values, and I did for a while, but I now prefer to have the fuel table contain the AFRs that I really want, and use the "Engine Load Compensation (MP)" table to fix issues that crop up in certain areas.