Statistics selection for sustainable cities datasets

[slumnitz]

Klick

Definition of information needed with statistics selection per Zaehlsprengel:

Displayed aggregated statistical information:

mean land surface temperature (2021): 39 degrees C - nice to have Existing GR area: 20 m2 - must Suitable GR area: 40 m2 - must Unused Potential Area for GR: 50 % - must to nice to have

displayed scatterplot is indicating where the Zeahlsprengel is compared to other Zaelsprengel of this city.

Updated as indicated in https://github.com/orgs/eurodatacube/projects/3/views/2?pane=issue&itemId=12099405

[lubojr]

@slumnitz @patrick-griffiths

As a first draft of summary statistics functionality, I have now enabled this aggregation of Green Roofs over ZSP. It is currently not possible for me to additionally configure it to get statistics after drawing a custom area from our geodb unfortunately, so this stays as a future enhancement.

Is the current aggregation like it is now meaningful ? I was not exposed to these data yet before.

variable	aggregated value
Mean Land Surface Temperature (2021)	41.2 degrees C
Total Roof area	17974 m²
Existing Green Roof area with a slope < 9 degree	1628 m²
Existing Green Roof area with a slope ≥ 9 and < 15 degree	2093 m²
Existing Green Roof area with a slope ≥ 15 and < 20 degree	0 m²
Unused Potential Area for Green Roof	79.30 %

The Unused potential Area is computed as 1 - (grpotare9 + grpotare15 + grpotare20) / roofArea) * 100;

Is there any meaningful aggregation also to be configured for Solar roofs?

[slumnitz]

Additionally, there was the wish to display the aggregation not only on Zeahlsprengel, but also on all aggregation levels (Bund, lavender, Nuts, lasso select).

TODO define content of different aggregation levels

[lubojr]

@slumnitz We have progressed on the multiselection topic - see current state on https://gtif-testing.eox.at/gtif_multiselect_groofs/ (still including "old" data for 4 cities only)

We have changed the chart x/y axis to display GRExisting vs GRPotential and when user selects multiple ZSPs, they get plotted on a scatterplot (label on hover is still broken - showing all points combined, I am aware). ZSPs now have a different color, while the color of a single ZSP matches the selection stroke on the map. Therefore the chart x/y axis are updated as per https://github.com/eurodatacube/eodash/issues/2008#issue-1577964186

But we thought that current visualisation needs quite a lot of interaction from the user, maybe it would be an easier option to let the user select one or more gemeindes and under the hood, this would fetch all data from ZSPs in those gemeindes (we would use the fact that ZSP IDs always contain the original gemeinde ID as part of it). In the chart, we would colorize all points for a single gemeinde in a same color.

Would that make sense for possible users? If not, we would need some guidance on the aggregation and "what we want to display on a chart". Thank you.

[Patrick1G]

@lubojr - very nice - but I agree: populating the scatterplot point by point is tedious. Selecting a Gemeinde with all ZSPs within would be much nicer...

[slumnitz]

@lubojr the original idea on the ZSP level was,

A) to fetch the data of the Gemeinde and plot all ZSP's from this Gemeinde, and for the user to then B) Select one or multiple ZSP that are then highlighted in colour on the scatterplot, to see where the ZSP is at in comparison to the Gemeinde.

In your example this would mean, the selection functionality stays. However, instead of only the selected ZSP the scatterplot would show all ZSP of the Gemeinde and color the selected ZSP.

Does this make sense to you and is it feasible, is also the question.

[lubojr]

@slumnitz @patrick-griffiths Amazing, thanks for a review. Yes, this should be feasible to do. Will let you know on progress.

[lubojr]

@slumnitz @patrick-griffiths Green roofs statistics of all zsp for a one or more gemeinde have been integrated to gtif.esa.int. Please have a look if okay.

The solar roofs currently do have neither aggregated statistics nor a correct scatter plot. How should we proceed there?

[lubojr]

After discussion of @santilland and @FranLeonelli - decided to implement statistics of solar roofs potential vs existing.

In data we have:

"pvpotentl": 1,
"pvexisting": 0,

[lubojr]

@FranLeonelli @Patrick1G @santilland

We have implemented the statistics table and gemeinde-level aggregation chart for solar roofs as well. It displays total area of potential solar roofs (x-axis) vs existing solar roof area (y-axis). We have also fixed the computation of the green roof area, so that it behaves the same as above.

Prepared a PR of a testing branch https://gtif-testing.eox.at/ui-panels-cat-roofs-stats-chart/. Please confirm if we can merge the branch and close this issue.

Thank you!

[santilland]

Ok from @Patrick1G and @FranLeonelli during biweekly, ready to merge. Look into adding a diagonal line information (create new issue?)

[santilland]

Look into adding ZSP name in feature selection overview

[santilland]

Reviewing the information available for Zählsprengel it seems that some might have actually a name, but most don't. "Eine Namensbezeichnung der Zählsprengel ist nicht allgemein erfolgt: Zählsprengel in Wien und anderen Städten werden nur durch ihre Nummern bezeichnet.[1] Wo es systematische Zählbezirke gibt, sind meist nur diese benannt." Neither the statistikal austria dataset nor the STAT Atlas have Zählsprengel names. Statistics for different administration levels is being discussed in this issue #2576 Added enhancement idea of diagonal line as new issue #2578 Closing this issue