I am trying to visualize a decision tree plot aka decision surface plot from a JSON containing coordinate mesh. It's a little tricky as the shapes are irregular and a small step size is needed to incrementally interpolate spatially by that small step size (say .02) until each zig zag has been accounted for. In my case, I have a matrix of 220 x 395 of purely x coordinates, and a y coordinate matrix of equal size, as well as an ID matrix for the predicted class (also governs color).
With these coordinate and ID values, I want to create a 2D visual of these irregular shapes using D3's path. You can see my best attempt on this block:
No errors are displaying, but I know something is amiss because nothing is being appended.
It should look like one of the following plots (without the scatter plot overlay, my data is only for the 2D regions):

For those familiar with Python, I'm using the numpy.meshgrid() call to get the coordinate data for the shapes, from which point I write to JSON. No knowledge of Python is required for my problem though I'm pretty sure, it's all about giving the D3 path the input it expect, which I'm confused about given my particular input.
My guess is I'm not accessing the JSON the correct way. Or I am not using the right line generating technique for the path to draw the lines at the JSON values. Can someone please explain what I can do to plot via D3 path correctly?
Thank you