Create plot_ghi_transposition.py example

[ericf900]

• I am familiar with the contributing guidelines
• New code is fully documented. Includes numpydoc compliant docstrings, examples, and comments where necessary.
• Pull request is nearly complete and ready for detailed review.
• Maintainer: Appropriate GitHub Labels and Milestone are assigned to the Pull Request and linked Issue.

Example showing transposition from GHi to POA, and comparing winter transposition gain to summer.

[kandersolar]

Hey @ericf900, cool example! This PR's version of the docs are built here, so you can check out how your example looks when it's live (the build takes a few minutes after you push an update): https://pvlib-python--933.org.readthedocs.build/en/933/

A couple notes:

• The docs build system only build files whose names start with "plot_", that's why your example isn't showing up like the others. If you rename the file and then push that commit, the docs will get rebuilt.
• Typically we strip out the spyder header (# -*- coding: utf-8 -*- / Created on... etc) before merging into pvlib
• You should add an entry about this example under the Documentation header and (if you're comfortable with it) your name to the Contributors section in the version 0.7.2 whatsnew file: https://github.com/pvlib/pvlib-python/blob/master/docs/sphinx/source/whatsnew/v0.7.2.rst

[kandersolar]

Also can you merge upstream/master into your branch? Some fixes were recently made to clean up the documentation build log and it would make reviewing this a little easier.

[ericf900]

@kanderso-nrel Thanks for the suggestions! I made those updates and updated the PR

[kandersolar]

Here are some suggestions. The :py:meth:... stuff will auto-link the functions to their documentation entries (see here for an example of what it looks like). Otherwise looks good to me!

[kandersolar]

very nice choice of location, +1

[kandersolar]

Suggested change

	solar_zenith=solar_position['zenith'],
	solar_zenith=solar_position['apparent_zenith'],

FYI I'd like someone else to confirm that using apparent solar position is appropriate before making this change

[cwhanse]

According to the parameters for each transposition model option in get_total_irradiance, apparent_zenith should be passed through, rather than zenith.

It's a different matter to verify that each model actually expects apparent_zenith rather than zenith. I'll look at the underlying references when I can find the opportunity.

[mikofski]

This example is really great! Thanks!

I have a minor suggestion, can you vertically stack the plots and share the x-axis?

fig, ax = plt.subplots(2, 1, sharex=True)  # stack plots (2, 1) and share the x axis
ax1, ax2 = ax  # or you can just use ax[0] instead of ax1, and ax[1] instead ax2, minor preference

I'm having a hard time reading the dates on the x-axis because they're a bit crowded, so I thought trying them vertical and sharing them might look nicer, but your call - but not a blocker for me - also fine as is, THANKS!

[kandersolar]

Unfortunately the indexes are 6 months apart so I don't think sharex=True will look good. Could convert the datetime index to a nicer string form with df.index.strftime("%H:%M"). Or could drop sharex=True, but then the x-label for the upper axes might overlap the bottom axes.

[ericf900]

Hi Mike and Kevin! I am looking at this now- I agree the x-axis is a bit cluttered. I think the benefit of keeping the plots side-by-side is that it highlights that, while there is not much of a gain for POA compared to GHI in the summer, overall irradiance is higher.

I'll make some edits now and push up the new version!

[mikofski]

You're both right! Sorry for the distraction! Great work

[cwhanse]

Thanks @ericf900

[cwhanse]

+1 for this note, answers a question often asked by newcomers to PV modeling

[cwhanse]

According to the parameters for each transposition model option in get_total_irradiance, apparent_zenith should be passed through, rather than zenith.

It's a different matter to verify that each model actually expects apparent_zenith rather than zenith. I'll look at the underlying references when I can find the opportunity.

[mikofski]

Eric, this is a really great contribution! Thanks so much. I really like the comments, super thorough and easy to read. Very good explanation of the process of generating clear-sky and transposing to POA. Awesome!

[mikofski]

you can use latex here for W/m^2 if you want, it might look a little bit nicer?

ax1.set_ylabel('Irradiance ($W/m^2$)')

I also added gridlines, but I noticed other examples don't have them, so just my personal preference, dk what the convention is.

[ericf900]

Thanks for the feedback Mike! Made these changes and am uploading a new version now :)

[wholmgren]

Thanks @ericf900 for the great example and @kanderso-nrel @mikofski @cwhanse for the reviews. I suggest merging as is.