Huh, this looks like the same thing I came up with about three years ago. I'd tracked my weight for a long time and only just started tracking calories, then a month or so in wondered if combining the two datasets would result in anything useful. After about six months there were some obvious patterns, and ever since a year in I've been able to control my weight however I want without much effort.
One graph I have that I don't see on there is average-calories-per-day on the X axis and weight-change-per-week on the Y axis (scatter plot). Sticking a linear regression on top told me all those online estimators for how many calories you burn in a day are just wrong, at least for me.
That scatter plot is a nice addition. One property of the Hacker's Diet is that the absolute numbers don't really matter, as long as the weighing and calorie balance are roughly consistent. What's relevant for the feedback loop is the weight slope, and the target can be something like losing 15% of weight, and a 10% calorie deficit might get you there in n months.