Re: [QGIS-Developer] Processing 3.0: Dropping non-QgsZonalStatistics zonal statistics algorithm

[Mathieu Pellerin]

This PR (https://github.com/qgis/QGIS/pull/4792) adds variance to
QgsZonalStatistics, and remove the duplicate processing zonal statistics
algorithm on the basis of a general consensus in this mail tread and more
importantly its author, Alexander Bruy, agreeing.

Thanks for the time spent discussing this.

On Wed, Jun 28, 2017 at 3:18 PM, Paolo Cavallini 
wrote:

> Il 28/06/2017 09:48, Nyall Dawson ha scritto:
>
> > We could just rename the option as "Majority (mode)" - does that work
> > for you? (or "Mode (majority)").
>
> fine for me; I just think mode is much more used in stats.
>
> >> Would it be possible to also add variance?
> >
> > Should be trivial to do - it's just the square of the standard deviation.
>
> sure - most users will find it convenient to have it ready made though.
>
> all the best, and thanks.
>
> --
> Paolo Cavallini - www.faunalia.eu
> QGIS & PostGIS courses: http://www.faunalia.eu/training.html
> https://www.google.com/trends/explore?date=all=IT=qgis,arcgis
>
___
QGIS-Developer mailing list
QGIS-Developer@lists.osgeo.org
List info: https://lists.osgeo.org/mailman/listinfo/qgis-developer
Unsubscribe: https://lists.osgeo.org/mailman/listinfo/qgis-developer

Re: [QGIS-Developer] Processing 3.0: Dropping non-QgsZonalStatistics zonal statistics algorithm

[Paolo Cavallini]

Il 28/06/2017 09:48, Nyall Dawson ha scritto:

> We could just rename the option as "Majority (mode)" - does that work
> for you? (or "Mode (majority)").

fine for me; I just think mode is much more used in stats.

>> Would it be possible to also add variance?
> 
> Should be trivial to do - it's just the square of the standard deviation.

sure - most users will find it convenient to have it ready made though.

all the best, and thanks.

-- 
Paolo Cavallini - www.faunalia.eu
QGIS & PostGIS courses: http://www.faunalia.eu/training.html
https://www.google.com/trends/explore?date=all=IT=qgis,arcgis
___
QGIS-Developer mailing list
QGIS-Developer@lists.osgeo.org
List info: https://lists.osgeo.org/mailman/listinfo/qgis-developer
Unsubscribe: https://lists.osgeo.org/mailman/listinfo/qgis-developer

Re: [QGIS-Developer] Processing 3.0: Dropping non-QgsZonalStatistics zonal statistics algorithm

[Nyall Dawson]

On 28 June 2017 at 16:23, Paolo Cavallini  wrote:

>
>
> Fine. Could we then rename from majority to mode? More understandable IMHO.

We could just rename the option as "Majority (mode)" - does that work
for you? (or "Mode (majority)").

> Would it be possible to also add variance?

Should be trivial to do - it's just the square of the standard deviation.

Nyall
___
QGIS-Developer mailing list
QGIS-Developer@lists.osgeo.org
List info: https://lists.osgeo.org/mailman/listinfo/qgis-developer
Unsubscribe: https://lists.osgeo.org/mailman/listinfo/qgis-developer

Re: [QGIS-Developer] Processing 3.0: Dropping non-QgsZonalStatistics zonal statistics algorithm

[Paolo Cavallini]

Il 28 giugno 2017 08:07:12 CEST, Nyall Dawson  ha 
scritto:
>On 28 June 2017 at 15:55, Mathieu Pellerin 
>wrote:
>>
>
>> Two statistical outputs (variance and mode) aren't available in
>QgsZonalStatistics-based alg, while two outputs (minority and majority)
>are missing from the alternate alg.
>
>Thanks!
>
>Mode = majority here, so both algs offer this value.
>
>Variance is trivial to calculate from the standard deviation, so
>that's not a significant loss.
>
>I guess the other unknowns are speed & robustness of the two
>algorithms. Is there any open bugs filed against either on hub?
>
>Nyall
>___
>QGIS-Developer mailing list
>QGIS-Developer@lists.osgeo.org
>List info: https://lists.osgeo.org/mailman/listinfo/qgis-developer
>Unsubscribe: https://lists.osgeo.org/mailman/listinfo/qgis-developer

Fine. Could we then rename from majority to mode? More understandable IMHO.
Would it be possible to also add variance?
Thanks.
-- 
Sent from mobile. Sorry for being short___
QGIS-Developer mailing list
QGIS-Developer@lists.osgeo.org
List info: https://lists.osgeo.org/mailman/listinfo/qgis-developer
Unsubscribe: https://lists.osgeo.org/mailman/listinfo/qgis-developer

Re: [QGIS-Developer] Processing 3.0: Dropping non-QgsZonalStatistics zonal statistics algorithm

[Nyall Dawson]

On 28 June 2017 at 15:55, Mathieu Pellerin  wrote:
>

> Two statistical outputs (variance and mode) aren't available in 
> QgsZonalStatistics-based alg, while two outputs (minority and majority) are 
> missing from the alternate alg.

Thanks!

Mode = majority here, so both algs offer this value.

Variance is trivial to calculate from the standard deviation, so
that's not a significant loss.

I guess the other unknowns are speed & robustness of the two
algorithms. Is there any open bugs filed against either on hub?

Nyall
___
QGIS-Developer mailing list
QGIS-Developer@lists.osgeo.org
List info: https://lists.osgeo.org/mailman/listinfo/qgis-developer
Unsubscribe: https://lists.osgeo.org/mailman/listinfo/qgis-developer

Re: [QGIS-Developer] Processing 3.0: Dropping non-QgsZonalStatistics zonal statistics algorithm

[Mathieu Pellerin]

Here's a table showing the overlap (and unique features) of both algs:


*QgsZonalStatistics alg* *Alternate alg*
Count x x
Sum x x
Mean x x
Median x x
StDev x x
Min x x
Max x x
Range x x
Minority x
Majority x
Variety / Unique x x
Variance (i.e. numpy.var)
x
Mode (I.e.scipy.stats.mode)
x

Two statistical outputs (variance and mode) aren't available in
QgsZonalStatistics-based alg, while two outputs (minority and majority) are
missing from the alternate alg.

I'm not an intense user of zonal stats either, and I've always used the
QgsZoneStatistics one.



On Wed, Jun 28, 2017 at 12:07 PM, Nyall Dawson 
wrote:

> On 28 June 2017 at 12:19, Mathieu Pellerin  wrote:
> > As part of the ongoing upgrade and cleanup done by the Nyall, we should
> take
> > the opportunity to get rid of one of the overlapping "Zonal Statistics"
> and
> > "Zonal Statistics (QGIS)" algorithms. The latter relies on QGIS'
> > QgsZonalStatistics.
> >
> > Yesterday, I've resurrected the "Zonal Statistics (QGIS)" algorithm, and
> in
> > doing so added progress feedback + cancellation support. Would anyone
> object
> > to:
> > a/ rename this algorithm to "Zonal Statistics"
> > b/ delete the former (and obsolete) "Zonal Statistics" algorithm
>
> I'm not a big user of zonal stats... but I'd like to see this
> duplication removed.
>
> Can you sum up what the differences between the tool algorithms are?
> Are they any downsides to dropping the other tool, or is the Zonal
> Stats (QGIS) alg better in every way?
>
> Nyall
>
___
QGIS-Developer mailing list
QGIS-Developer@lists.osgeo.org
List info: https://lists.osgeo.org/mailman/listinfo/qgis-developer
Unsubscribe: https://lists.osgeo.org/mailman/listinfo/qgis-developer

Re: [QGIS-Developer] Processing 3.0: Dropping non-QgsZonalStatistics zonal statistics algorithm

[Nyall Dawson]

On 28 June 2017 at 12:19, Mathieu Pellerin  wrote:
> As part of the ongoing upgrade and cleanup done by the Nyall, we should take
> the opportunity to get rid of one of the overlapping "Zonal Statistics" and
> "Zonal Statistics (QGIS)" algorithms. The latter relies on QGIS'
> QgsZonalStatistics.
>
> Yesterday, I've resurrected the "Zonal Statistics (QGIS)" algorithm, and in
> doing so added progress feedback + cancellation support. Would anyone object
> to:
> a/ rename this algorithm to "Zonal Statistics"
> b/ delete the former (and obsolete) "Zonal Statistics" algorithm

I'm not a big user of zonal stats... but I'd like to see this
duplication removed.

Can you sum up what the differences between the tool algorithms are?
Are they any downsides to dropping the other tool, or is the Zonal
Stats (QGIS) alg better in every way?

Nyall
___
QGIS-Developer mailing list
QGIS-Developer@lists.osgeo.org
List info: https://lists.osgeo.org/mailman/listinfo/qgis-developer
Unsubscribe: https://lists.osgeo.org/mailman/listinfo/qgis-developer

Re: [QGIS-Developer] Processing 3.0: Dropping non-QgsZonalStatistics zonal statistics algorithm

[Paolo Cavallini]

Il 28/06/2017 04:19, Mathieu Pellerin ha scritto:
> As part of the ongoing upgrade and cleanup done by the Nyall, we should
> take the opportunity to get rid of one of the overlapping "Zonal
> Statistics" and "Zonal Statistics (QGIS)" algorithms. The latter relies
> on QGIS' QgsZonalStatistics.
> 
> Yesterday, I've resurrected the "Zonal Statistics (QGIS)" algorithm, and
> in doing so added progress feedback + cancellation support. Would anyone
> object to:
> a/ rename this algorithm to "Zonal Statistics"
> b/ delete the former (and obsolete) "Zonal Statistics" algorithm

Merci Mathieu.
Which alg are you talking about? I can only remember of a single tool.
All the best.

-- 
Paolo Cavallini - www.faunalia.eu
QGIS & PostGIS courses: http://www.faunalia.eu/training.html
https://www.google.com/trends/explore?date=all=IT=qgis,arcgis
___
QGIS-Developer mailing list
QGIS-Developer@lists.osgeo.org
List info: https://lists.osgeo.org/mailman/listinfo/qgis-developer
Unsubscribe: https://lists.osgeo.org/mailman/listinfo/qgis-developer