I learned how to arrange labels on PyPlot charts yesterday.
I've mentioned that the scales on the ham radio QSO elevation profiles weren't entirely comparable to one another. Due to the variety off hills, mountain ridges, and coasts the kids and I make POTA activations on, the vertical scales on the elevation chart y axes range from three meters in total to dozens of meters. What's important (and interesting) about the various sites though isn't the sheer number of meters of descent or ascent, but rather, the angle of the terrain with respect to level. That angle gets subtracted from the natural radiation angle of the dipole that houses Project TouCans and, theoretically, extends the distance that TouCans can transmit before bouncing off the F2 layer of the ionosphere.
Consequently, rather than determining a scaling algorithm to make all the elevation profile graphs returned for the various QSOs comparable, I elected to simply calculate the slope of the terrain over the first two wavelengths, (using 2 20 meter wavelengths in these examples since Project TouCans is a 20 meter ham radio.) That just gave me the job of adding a label to each graph (after using scipy.stats to calculate the slope.) Getting the label to the correct spot on the graph took a little more work than I initially expected.
ChatGPT's original suggestion was to use:
plt.text(0.5, 50, slope_st, fontsize=12, ha='center')
That gave back graphs with a slope message that scootched around.
It was centered, but centered on what? Turns out, the text was centered on the coordinates specified in the argument, and those coordinates were literally tied to the data values. So, centering wasn't the easy way for me. I moved to 'left' for the ha argument. Sometimes things looked good:
But also? Sometimes, the label just disappeared:
After a bit of puzzling, I figured out that the data coordinate for 50, specified in the line of code above, for the y coordinate was well off the screen in the plot above.
Ultimately I arrived at the correct—for me—answer. I set ha, the horizontal alignment field, to 'left' rather than center. That caused the text to place it's left-most point at the coordinates specified in the method call: two tenths of the way to the right across the plot. I also used the transform argument:
transform=ax.transAxes
Which made the coordinates relative to the graph's axes, and scaled from 0,0 (lower left corner), to 1,1, (upper right corner.) I had to add a line of code to get the axes object before sending in the transform argument. So I wound up with:
ax = plt.gca()
Followed by
plt.text(0.2, 0.1, slope_st, fontsize=12, ha='left', transform=ax.transAxes)
Which gave me a label at the same spot on every plot independent of the length of the label or the values on either data axis:
Comments
Post a Comment
Please leave your comments on this topic: