Feature/tutorial review impact calculation

[sarah-hlsn]

This PR updates the Impact Calculation tutorial

Changes proposed in this PR:

• updated & fixed errors in code
• removed visualisation section at the end (the basic visualisations for impacts are covered throughout the tutorial; if we want to retain information on fancier plots I would suggest moving that to a dedicated tutorial, rather than weighing this one down)
• added a section about centroid assignment
• light edits to the text for clarity

Note: I had to rerun a lot of parts of this notebook due to code changes througout.

PR Author Checklist

• Read the Contribution Guide
• Correct target branch selected (if unsure, select develop)
• Descriptive pull request title added
• Source branch up-to-date with target branch
• Documentation updated
• Tests updated
• Tests passing
• No new linter issues
• Changelog updated

PR Reviewer Checklist

• Read the Contribution Guide
• CLIMADA Reviewer Checklist passed
• Tests passing
• No new linter issues

[emanuel-schmid]

This is not 100% correct, I'd say. 🧐

Exposures are geopandas dataframes with at least columns ‘latitude’, ‘longitude’ and ‘value’ of exposures. They are either defined as a series of (latitude/longitude) points or as a raster of (latitude/longitude) points. Fundamentally, this changes nothing for the impact computations. Note that for larger number of points, consider using a raster which might be more efficient (computationally). For a low number of points, avoid using a raster if this adds a lot of exposures values equal to 0.

Rather an Exposures contains a GeoDataFrame. And this one doesn't have any 'latitude'/'longitude' columns anymore.

[sarah-hlsn]

Thanks for catching that @emanuel-schmid! I'll edit accordingly.

[peanutfun]

@sarah-hlsn Not to worry. The test fails because since the v6.1 release, CLIMADA does not support Python 3.10 anymore. But the pipeline still tries to build an environment based on 3.10, so it's expected to fail. The error occurs now because develop was merged into this branch after the release.

[sarah-hlsn]

Great, thanks for checking @peanutfun !

[NicolasColombi]

Hi Sarah, the PR looks good to me!

[spjuhel]

Sorry, I thought this review was submitted already a long time ago.

Most important is the comment on the intensities/impact values in the plot.

[spjuhel]

• "assingning" typo
• Probably unrelated to the tutorial, but I wonder why are the hazard intensities/impacts so different for the "User defined point exposure and Tropical Cyclone hazard". The intensities look wrong, actually.

[spjuhel]

Great job!

I'm really curious/concerned with the intensities being between 0 and 1 in the point exposure example. This does not look to be related to your changes, but something is wrong somewhere. If you can briefly investigate, that would be wonderful.

[luseverin]

Hey, regarding the issues with the plot of the intensities, this was directly related to the plotting, i.e. the wrong collection was selected to be plotted:
ax_pnt.get_figure().colorbar( ax_pnt.collections[0], fraction=0.0175, pad=0.08 )
which was probably plotting some unrelated elements of the map. Selecting ax_pnt.collections[2] points to the correct data.