[Qgis-psc] First sample outputs of telemtry work
Jonathan Moules
jonathan-lists at lightpear.com
Thu May 12 12:28:36 PDT 2022
Hi Tim,
Thanks for the login credentials. I did some fiddling with the versions:
* Some SQL to get clean QGIS versions
* QGIS Version 30-day graph now uses this (Major.Minor version only).
* 20 most used full-versions (Major.Minor.Patch) for the past week
(table) - so we can see what users are using now(ish)
* History of QGIS version use as an area graph - very cool - got it by
fiddling around with things I didn't understand. :-D
* Stacked bar graph of versions. Basically the same as history-of-qgis
but different representation (not sure as to the value of this one;
happy to delete if no-one has a use for it).
You can see them all a little over half-way down on the dashboard. -
http://feed.qgis.org/metabase/public/dashboard/df81071d-4c75-45b8-a698-97b8649d7228
Cheers,
Jonathan
On 2022-05-11 00:39, Tim Sutton wrote:
> Hi Jonathan
>
> Thanks for your email, see comments below:
>
>
>
> On Tue, May 10, 2022 at 9:05 PM Jonathan Moules
> <jonathan-lists at lightpear.com> wrote:
>
> Hi Tim,
>
> This looks good; I love a good dashboard. Some thoughts & bugs:
>
> * Do we have telemetry for QGIS Server?
>
> No - I don't think we will add this.....or at least I don't have any
> current plans.
>
>
> * The numbers on "*QGIS Open events on Windows per day - last
> yea**r*" seem far too low. "*QGIS Daily Opens - last 30 days*" is
> showing half a million users a day but for windows is just ~50k a
> day, this despite windows clearly being 90+% of users per the
> platform chart (well, one of those platform charts...).
>
> Only QGIS >= 3.24 reports the platform. This figure will self correct
> as more people move to more recent builds.
>
> * Alternately, the pie-chart next to it: "*QGIS Daily Opens by
> Platform - last 30 days*" is showing 86% as being "other"
> platforms where "others" on the chart to the left of it for what
> appears to be trying to show the same thing is only a tiny
> percentage (< 6%?).
>
> Again this is explained by the lack of platform reporting in QGIS >
> 3.24 (see the little info popups on each chart).
>
>
> * "*QGIS Daily Opens by Version - last 30 days*" - is showing some
> of the version numbers as integers and some as floats.
>
>
> Yes I haven't figured this out yet....it could be bogus User Agents,
> custom builds of QGIS.....something else?
>
> * "*QGIS Top Movers - Yesterday*" - If I sum these up I get about
> 100,000. Given the total number of users yesterday per the
> day-chart was only about 25k higher than a Monday average, I get
> the feeling this is averaging across all-time, meaning weekends
> will always show a low number and weekdays will always show a high
> number (as today). I'd suggest it should probably use the average
> for that-day, so it's comparing like-for-like.
>
> Yeah this is a bit of an experiment. It compares the amount of opens
> yesterday to the average daily opens across all time. I don't know how
> informative it is, in future I may just remove it.
>
> * "*QGIS Total Number of Opens in this Current Month*" at the top
> is wrong - it's identical to "last 30 days" to its left. Also the
> very bottom shows the number should be 3.5 million not 10 million.
>
>
> Thanks, fixed now.
>
> * "*QGIS Total Number of Opens in this Current Month*" at the top
> says 499,262, but the graph (*QGIS Daily Opens - last 30 days*)
> shows 500,359 for yesterday if I mouse over it.
>
>
> I'm not sure why they are different - will need to poke around some more.
>
> Assuming it's SQL behind the scenes, I may be able to take a stab
> at it if you want.
>
> Yes I would welcome your help. I will send you some credentials, but
> please tray to avoid running any queries that place a high load on the
> server.
>
> Regards
>
> Tim
>
>
> Cheers,
>
> Jonathan
>
>
> On 2022-04-29 22:35, Tim Sutton wrote:
>> Hi Andreas
>>
>> On Fri, Apr 29, 2022 at 5:00 PM Andreas Neumann
>> <andreas at qgis.org> wrote:
>>
>> Hi Tim,
>>
>>
>> See also my answers to Régis for more context. More comments
>> below. Note that I share the dashboard publicly so look here
>> http://feed.qgis.org/metabase/public/dashboard/df81071d-4c75-45b8-a698-97b8649d7228
>>
>> especially at the little (i) indicators which I tried to populate
>> with more defail.
>>
>>
>> Looks really cool - thanks for working on this! I kind of
>> expected the dominance of Windows as an operating system, but
>> I am surprised to see one particular Windows (Windows 10)
>> take such a large share. Many government agencies or
>> companies that I know still use older Windows versions.
>>
>>
>> I guess MS has been working hard to make the upgrade process
>> smoother which is IMHO a good thing from a security point of view.
>>
>>
>> I also wonder what hides behind some of the quite large bars
>> in the category "Other"?
>>
>>
>> We can break them out if needed.
>>
>>
>>
>>
>> When it comes to QGIS versions, I think we should better
>> aggregate only by major versions and for now ignore the minor
>> versions. This large fragmentation into minor/minor versions
>> seems to lead to this gigantic large bar of "Other".
>>
>> Yes it is on my todo list to aggregate all 3xxyy into 3xx so they
>> show as one bucket.
>>
>>
>> Question: is there any chance to put the absolute numbers in
>> relation with the rough population size of a country. It is
>> only natural that large countries like France, Germany, USA
>> and Brazil open QGIS more often than smaller countries. But
>> put in relation with population size, the pattern could look
>> quite different.
>>
>>
>> The first map already does this. The second one shows absolute
>> numbers per territory regardless of population.
>>
>>
>>
>> My cartography teachers told me that it is absolutely
>> forbidden to use absolute numbers in choropleth maps -
>> ABSOLUTELY FORBIDDEN ! Numbers always need to be put in
>> relation with the size (area and/or population) of a country.
>>
>>
>> He is probably right but I think the world of BI (Business
>> Intelligence) has largely ignored him and the world is now
>> littered with maps like this. I still think it is interesting to
>> see the absolute numbers by country.
>>
>>
>> Thanks a lot for your work to all people involved!
>>
>>
>> No problem, and again let me extend the invitation to you to go
>> onto the backend and help improve the dashboard, this is just my
>> first chicken scratch. I think the real value will be to see the
>> relative breakdown of platforms, versions countries and the
>> changes to these over time so we can identify trends. In terms of
>> absolute numbers I am still unsure if it tells us a real picture
>> of how many users we have.
>>
>> Regards
>>
>> Tim
>>
>>
>> Andreas
>>
>>
>>
>> On Fri, 29 Apr 2022 at 15:57, Tim Sutton <tim at kartoza.com> wrote:
>>
>> Hi PSC
>>
>> I just wanted to give you some feedback on the work we
>> have been doing to use anonymised telemetry data from the
>> feed to try to get a handle on the number, geography and
>> platform choice of our user base. There are detailed
>> notes in the image below which should explain things.
>> Currently it has only been running about 2 weeks so the
>> year and month charts at the end look the same, but you
>> can expect these to diverge over time.
>>
>> If you have other things you would like to report on with
>> the available data, please pop me a note. I still want to
>> aggregate the platform better to lump together Win, Mac,
>> Linux and other instead of the more granular breakdown we
>> get right now. I will do so when I get a chance.
>>
>> I do not plan to make the link to the live dashboard
>> public as I don't want to put load on the feed server,
>> but if you wish to access the dashboard, pop me a note
>> and I will add you.
>>
>> https://user-images.githubusercontent.com/178003/165956926-a30c6c96-3b6b-4895-8afb-97e9ba0bd28c.png
>>
>>
>> Regards
>>
>> Tim
>> --
>> ------------------------------------------------------------------------------------------
>>
>> Tim Sutton
>> Visit http://kartoza.com to find out about open source:
>> * Desktop GIS programming services
>> * Geospatial web development
>> * GIS Training
>> * Consulting Services
>> Tim is a member of the QGIS Project Steering Committee
>> -------------------------------------------------------------------------------------------
>> _______________________________________________
>> Qgis-psc mailing list
>> Qgis-psc at lists.osgeo.org
>> https://lists.osgeo.org/mailman/listinfo/qgis-psc
>>
>>
>>
>> --
>>
>> --
>> Andreas Neumann
>> QGIS.ORG <http://QGIS.ORG> board member (treasurer)
>>
>>
>>
>> --
>> ------------------------------------------------------------------------------------------
>>
>> Tim Sutton
>> Visit http://kartoza.com to find out about open source:
>> * Desktop GIS programming services
>> * Geospatial web development
>> * GIS Training
>> * Consulting Services
>> Tim is a member of the QGIS Project Steering Committee
>> -------------------------------------------------------------------------------------------
>>
>> _______________________________________________
>> Qgis-psc mailing list
>> Qgis-psc at lists.osgeo.org
>> https://lists.osgeo.org/mailman/listinfo/qgis-psc
>
>
>
> --
> ------------------------------------------------------------------------------------------
>
> Tim Sutton
> Visit http://kartoza.com to find out about open source:
> * Desktop GIS programming services
> * Geospatial web development
> * GIS Training
> * Consulting Services
> Tim is a member of the QGIS Project Steering Committee
> -------------------------------------------------------------------------------------------
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/qgis-psc/attachments/20220512/d6e781bc/attachment-0001.htm>
More information about the Qgis-psc
mailing list