Hi,
our current definiton of „active editor” [1]:
An 'active editor' is a registered (and signed in) person (not known as a bot) who makes 5 or more edits in any month in mainspace on countable pages.
is centered around months. That's good.
However, as we are seeing requests to produce daily graphs: How to interpret the above definition in terms of active editors for a given /day/?
Especially: How to do it in a way that blends nicely with the current month-based definition?
Best regards, Christian
P.S.: I've seen code in our repos that just looks for edits of the last 30 days. That sounds nice. But if I am doing 3 edits on 2013-07-01, and another 3 on 2013-07-31, I would not be considered active editor by this daily approach for any day. However, I'd be an active editor for July using [1] :-/
[1] https://www.mediawiki.org/wiki/Analytics/Metric_definitions#Active_editor
We could use a moving 30 day window. If we update the definition slightly to say that an active editor is an editor that has made at least 5 edits in the last 30 days (effectively what's happening now), we can generate that every day.
On Wed, Sep 25, 2013 at 10:36 AM, Christian Aistleitner < christian@quelltextlich.at> wrote:
Hi,
our current definiton of „active editor” [1]:
An 'active editor' is a registered (and signed in) person (not known as a bot) who makes 5 or more edits in any month in mainspace on countable pages.
is centered around months. That's good.
However, as we are seeing requests to produce daily graphs: How to interpret the above definition in terms of active editors for a given /day/?
Especially: How to do it in a way that blends nicely with the current month-based definition?
Best regards, Christian
P.S.: I've seen code in our repos that just looks for edits of the last 30 days. That sounds nice. But if I am doing 3 edits on 2013-07-01, and another 3 on 2013-07-31, I would not be considered active editor by this daily approach for any day. However, I'd be an active editor for July using [1] :-/
[1] https://www.mediawiki.org/wiki/Analytics/Metric_definitions#Active_editor
-- ---- quelltextlich e.U. ---- \ ---- Christian Aistleitner ---- Companies' registry: 360296y in Linz Christian Aistleitner Gruendbergstrasze 65a Email: christian@quelltextlich.at 4040 Linz, Austria Phone: +43 732 / 26 95 63 Fax: +43 732 / 26 95 63 Homepage: http://quelltextlich.at/
Analytics mailing list Analytics@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/analytics
+1 to 30 day sliding window.
On Wed, Sep 25, 2013 at 8:55 AM, Aaron Halfaker aaron.halfaker@gmail.comwrote:
We could use a moving 30 day window. If we update the definition slightly to say that an active editor is an editor that has made at least 5 edits in the last 30 days (effectively what's happening now), we can generate that every day.
On Wed, Sep 25, 2013 at 10:36 AM, Christian Aistleitner < christian@quelltextlich.at> wrote:
Hi,
our current definiton of „active editor” [1]:
An 'active editor' is a registered (and signed in) person (not known as a bot) who makes 5 or more edits in any month in mainspace on countable pages.
is centered around months. That's good.
However, as we are seeing requests to produce daily graphs: How to interpret the above definition in terms of active editors for a given /day/?
Especially: How to do it in a way that blends nicely with the current month-based definition?
Best regards, Christian
P.S.: I've seen code in our repos that just looks for edits of the last 30 days. That sounds nice. But if I am doing 3 edits on 2013-07-01, and another 3 on 2013-07-31, I would not be considered active editor by this daily approach for any day. However, I'd be an active editor for July using [1] :-/
[1] https://www.mediawiki.org/wiki/Analytics/Metric_definitions#Active_editor
-- ---- quelltextlich e.U. ---- \ ---- Christian Aistleitner ---- Companies' registry: 360296y in Linz Christian Aistleitner Gruendbergstrasze 65a Email: christian@quelltextlich.at 4040 Linz, Austria Phone: +43 732 / 26 95 63 Fax: +43 732 / 26 95 63 Homepage: http://quelltextlich.at/
Analytics mailing list Analytics@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/analytics
Analytics mailing list Analytics@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/analytics
We could use a moving 30 day window. If we update the definition slightly
to say that an active editor is an editor that has made at least 5 edits in the last 30 days (effectively what's happening now), we can generate that every day.
Depending on what the context is for the chart that may confuse people. Suppose we have a vast but very transient influx of editors, say tenfold as much as usual on a smaller wiki, say after a national celebrity confesses she edits Wikipedia. Now in the extreme case where all new editors stay for just one day, that would still yield a high plateau in the chart for 30 days. The same of course would happen in monthly stats. But there people expect this kind of behavior. In daily stats now so much.
A small window of say three days would dampen this effect. Yes it'll be apples and oranges with monthly stats but that may be unavoidable.
Erik
From: analytics-bounces@lists.wikimedia.org [mailto:analytics-bounces@lists.wikimedia.org] On Behalf Of Aaron Halfaker Sent: Wednesday, September 25, 2013 5:55 PM To: A mailing list for the Analytics Team at WMF and everybody who has an interest in Wikipedia and analytics. Subject: Re: [Analytics] "Active editor" for a given day (instead of month)
We could use a moving 30 day window. If we update the definition slightly to say that an active editor is an editor that has made at least 5 edits in the last 30 days (effectively what's happening now), we can generate that every day.
On Wed, Sep 25, 2013 at 10:36 AM, Christian Aistleitner christian@quelltextlich.at wrote:
Hi,
our current definiton of "active editor" [1]:
An 'active editor' is a registered (and signed in) person (not known as a bot) who makes 5 or more edits in any month in mainspace on countable pages.
is centered around months. That's good.
However, as we are seeing requests to produce daily graphs: How to interpret the above definition in terms of active editors for a given /day/?
Especially: How to do it in a way that blends nicely with the current month-based definition?
Best regards, Christian
P.S.: I've seen code in our repos that just looks for edits of the last 30 days. That sounds nice. But if I am doing 3 edits on 2013-07-01, and another 3 on 2013-07-31, I would not be considered active editor by this daily approach for any day. However, I'd be an active editor for July using [1] :-/
[1] https://www.mediawiki.org/wiki/Analytics/Metric_definitions#Active_editor
-- ---- quelltextlich e.U. ---- \ ---- Christian Aistleitner ---- Companies' registry: 360296y in Linz Christian Aistleitner Gruendbergstrasze 65a Email: christian@quelltextlich.at 4040 Linz, Austria Phone: +43 732 / 26 95 63 tel:%2B43%20732%20%2F%2026%2095%2063 Fax: +43 732 / 26 95 63 tel:%2B43%20732%20%2F%2026%2095%2063 Homepage: http://quelltextlich.at/ ---------------------------------------------------------------
_______________________________________________ Analytics mailing list Analytics@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/analytics
Someday people will want edits by the hour, broken down by country and age group ;-) Seriously too much granularity can makes us focus on noise rather than trends. In the report card since we changed from 1 to 3 years as default charting period, we spend less time overanalyzing relatively small outliers.
We could determine 'normalized' daily counts by calculating for historic months ratio (monthly 5+)/(avg daily 1+), and scaling daily counts with this factor but it feels a bit artificial to me.
The alternative is either unscaled 1+ or 5+, and emphasize daily and monthly counts are not related. Same issue we have with daily vs monthly unique visitors (from comScore). We chose to ignore daily UV's.
Erik
-----Original Message----- From: analytics-bounces@lists.wikimedia.org [mailto:analytics-bounces@lists.wikimedia.org] On Behalf Of Christian Aistleitner Sent: Wednesday, September 25, 2013 5:36 PM To: analytics@lists.wikimedia.org Subject: [Analytics] "Active editor" for a given day (instead of month)
Hi,
our current definiton of „active editor” [1]:
An 'active editor' is a registered (and signed in) person (not known as a bot) who makes 5 or more edits in any month in mainspace on countable pages.
is centered around months. That's good.
However, as we are seeing requests to produce daily graphs: How to interpret the above definition in terms of active editors for a given /day/?
Especially: How to do it in a way that blends nicely with the current month-based definition?
Best regards, Christian
P.S.: I've seen code in our repos that just looks for edits of the last 30 days. That sounds nice. But if I am doing 3 edits on 2013-07-01, and another 3 on 2013-07-31, I would not be considered active editor by this daily approach for any day. However, I'd be an active editor for July using [1] :-/
[1] https://www.mediawiki.org/wiki/Analytics/Metric_definitions#Active_editor
-- ---- quelltextlich e.U. ---- \ ---- Christian Aistleitner ---- Companies' registry: 360296y in Linz Christian Aistleitner Gruendbergstrasze 65a Email: christian@quelltextlich.at 4040 Linz, Austria Phone: +43 732 / 26 95 63 Fax: +43 732 / 26 95 63 Homepage: http://quelltextlich.at/ ---------------------------------------------------------------
Hi Christian,
thanks for starting this thread. I believe the discussion is too important to be held on a mailing list and I suggest we start a wiki page discussing pros and cons of different proposals.
I have my own list of concerns with the canonical TAE definition (issues with how we define content namespaces, countable pages and the discounting of activity on deleted pages, lack of metrics focused on registration date) which I'd like to discuss. Since we're having this conversation I'd rather have it in a place where we can assess the strength of each proposal. How does that sound?
We also have stubs for user class definitions here [1] and it'd be great to start revamping that page.
Dario
[1] https://meta.wikimedia.org/wiki/Research:Metrics#User_classes
On Sep 25, 2013, at 9:30 AM, Erik Zachte ezachte@wikimedia.org wrote:
Someday people will want edits by the hour, broken down by country and age group ;-) Seriously too much granularity can makes us focus on noise rather than trends. In the report card since we changed from 1 to 3 years as default charting period, we spend less time overanalyzing relatively small outliers.
We could determine 'normalized' daily counts by calculating for historic months ratio (monthly 5+)/(avg daily 1+), and scaling daily counts with this factor but it feels a bit artificial to me.
The alternative is either unscaled 1+ or 5+, and emphasize daily and monthly counts are not related. Same issue we have with daily vs monthly unique visitors (from comScore). We chose to ignore daily UV's.
Erik
-----Original Message----- From: analytics-bounces@lists.wikimedia.org [mailto:analytics-bounces@lists.wikimedia.org] On Behalf Of Christian Aistleitner Sent: Wednesday, September 25, 2013 5:36 PM To: analytics@lists.wikimedia.org Subject: [Analytics] "Active editor" for a given day (instead of month)
Hi,
our current definiton of „active editor” [1]:
An 'active editor' is a registered (and signed in) person (not known as a bot) who makes 5 or more edits in any month in mainspace on countable pages.
is centered around months. That's good.
However, as we are seeing requests to produce daily graphs: How to interpret the above definition in terms of active editors for a given /day/?
Especially: How to do it in a way that blends nicely with the current month-based definition?
Best regards, Christian
P.S.: I've seen code in our repos that just looks for edits of the last 30 days. That sounds nice. But if I am doing 3 edits on 2013-07-01, and another 3 on 2013-07-31, I would not be considered active editor by this daily approach for any day. However, I'd be an active editor for July using [1] :-/
[1] https://www.mediawiki.org/wiki/Analytics/Metric_definitions#Active_editor
-- ---- quelltextlich e.U. ---- \ ---- Christian Aistleitner ---- Companies' registry: 360296y in Linz Christian Aistleitner Gruendbergstrasze 65a Email: christian@quelltextlich.at 4040 Linz, Austria Phone: +43 732 / 26 95 63 Fax: +43 732 / 26 95 63 Homepage: http://quelltextlich.at/
Analytics mailing list Analytics@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/analytics