Does this site track how much time you are logged in? That could make for some very depressing stats haha
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Top 50 Posters Extrapolated, January 2019
- Thread starter chown33
- Start date
- Sort by reaction score
Oh man, I’m not sure I want to see that statistic.
If I have 6 oranges in one hand, and four oranges in the other what have I got asked my teacher.
Big hands?
You have $1000.
Your ex texts asking for $500.
Your kids text asking for $300.
How much money do you have?
Correct answer: $1000 and several unread text messages.
Graphs of Post Counts
Based on measurements taken 03 April 2019 by chown33.
Info About All the Graphs
The horizontal (X) axes represent dates, and are labeled as YYYY-MM-DD.
The vertical (Y) axes are automatically determined, and are specific to each graph. The upper and lower limits of a graph are based on the range of post-counts or post-rates in the subset of graphed data. This means that the steepness of line slopes can't be compared between different graphs.
Colors are also automatically assigned, from a rotating pool of 12 colors. This can make color-to-name conversion ambiguous when there are more than 12 members graphed.
Use each graph's legend or key on the right to disambiguate duplicate colors. The legend is arranged in the same order as the right-most points of each graph. For example, in Graph 1, the upper red line in Group 1 is maflynn (top of legend), and the lower red line is someone else.
A dashed line is automatically assigned to inactive users (no posts in the latest measurement period), and a solid line to active users (at least 1 post in latest period). To reduce the clutter, the only inactive user typically shown is @mad jew.
Graphs 1: Active Top 50 Posters - 2 Groups, 1 Composite
These graphs represent Active Top 50 Posters, i.e. those with the highest post counts. @mad jew is the only inactive member shown, for historical reasons.
Member @eyoungren appears in both Group graphs, with different line slopes, showing that slopes between graphs aren't comparable.
Graphs 2: Weekly Post Growth - 3 Groups, 1 Composite
These graphs represent Prolific Posters, i.e. those with the highest post rates, regardless of their post count.
The post count of each member shown is sampled on the same day each week. The growth from the 1st sampled date is then drawn on the graph. Thus, the first value graphed for 2019-01-13 is the increase in post count for the week ending on 2019-01-13.
Remarks
Based on measurements taken 03 April 2019 by chown33.
Info About All the Graphs
The horizontal (X) axes represent dates, and are labeled as YYYY-MM-DD.
The vertical (Y) axes are automatically determined, and are specific to each graph. The upper and lower limits of a graph are based on the range of post-counts or post-rates in the subset of graphed data. This means that the steepness of line slopes can't be compared between different graphs.
Colors are also automatically assigned, from a rotating pool of 12 colors. This can make color-to-name conversion ambiguous when there are more than 12 members graphed.
Use each graph's legend or key on the right to disambiguate duplicate colors. The legend is arranged in the same order as the right-most points of each graph. For example, in Graph 1, the upper red line in Group 1 is maflynn (top of legend), and the lower red line is someone else.
A dashed line is automatically assigned to inactive users (no posts in the latest measurement period), and a solid line to active users (at least 1 post in latest period). To reduce the clutter, the only inactive user typically shown is @mad jew.
Graphs 1: Active Top 50 Posters - 2 Groups, 1 Composite
These graphs represent Active Top 50 Posters, i.e. those with the highest post counts. @mad jew is the only inactive member shown, for historical reasons.
Member @eyoungren appears in both Group graphs, with different line slopes, showing that slopes between graphs aren't comparable.
Graphs 2: Weekly Post Growth - 3 Groups, 1 Composite
These graphs represent Prolific Posters, i.e. those with the highest post rates, regardless of their post count.
The post count of each member shown is sampled on the same day each week. The growth from the 1st sampled date is then drawn on the graph. Thus, the first value graphed for 2019-01-13 is the increase in post count for the week ending on 2019-01-13.
Remarks
- The Group graphs are less cluttered and should be easier to follow than the Composite graphs. The Composites are useful mainly to see everyone's relative counts or rates on the same scale.
- @Scepticalscribe shows a classic deceleration curve starting at 2019-03-03. She mentioned this in a post above, and here's the weekly picture of it.
Very interesting. I can see a dip for the week I was at a trade show. Didn't have much free time.
Thanks for 'fun with graphs'!
Exactly.
Mind you, now that readings are taken on a weekly basis (rather than a monthly, three monthly or six monthly basis), it seems to be possible to focus very closely on a member's posting habits.
In truth, there is a very close relation between my posting habits when I am at home, and when I am posted abroad.
The fact that I was responsible for ordering, organising and running my mother's life whenever I was at home, (and we had an EPA in place, obtained from the High Court), meant that my life revolved around her needs, and that, in turn, meant a lot of time in front of the computer screen, as normal social activities were - to a large extent - pretty much curtailed, or suspended, or sometimes plans had to be cancelled at short notice if a domestic emergency happened.
@Apple fanboy will understand this, but domestic circumstances meant that face-to-face meeting with friends had become an annual thing - even, at times, a biennial occurrence; instead, most contact with friends took place over the phone, or text, or by email.
As always, an exceedingly interesting read, and thanks for your work, @chown33.
Last edited:
Agreed. It’s often my only social outlet week to week.
I used it for practical reasons, namely sorting. The scripts I use to produce the tables and graphs are combination of bash and awk. It's much simpler to work with strings that collate properly without conversion.
I do convert the date strings into a "days since Unix epoch" count that's used to determine the length of the intervals between dates. Without that, posts per day would be wrong.
I wasn’t joking. I use ISO when signing things by hand and a date is necessary. It just works for me.
(When ornery I use julian.)
When I'm not sure if someone is joking or not, I often just state the simple facts. Sometimes this helps to clarify things, but every once in a while I get a somewhat angry response like, "Are you being a smart ass?". My honest answer is "Just stating facts."
By hand, my habit is to go with "DD Month YYYY" where "Month" is an abbrev of the month name.
When I'm very ornery I use the French Republican Calendar, or my personal approximation of it. At least that way, if someone questions whether I'm being a smart ass I can answer "Oui".
Some of the post-count growth lines look almost horizontal, including mine. That effect would be lessened if the users with the highest counts, like maflynn, weren't on the same graph, thereby setting the vertical scale.
I like how I'm somehow still in the 30's even though I post about 15 times a year anymore.
I don't think it would make that much difference. I've made a few example graphs with higher-count users successively omitted, and your line stays pretty flat in all of them.
First though, let's see a graph with only the Top 7 active users:
You, GGJStudios, and Eidorian are all fairly flat: nearly horizontal and with few bends. The flattest of all is mad jew, who's a perfectly flat line, since he's inactive.
The high-rate posters are maflynn, C DM and ScepticalScribe. ScepticalScribe starts below your line and ends above it, as does C DM.
Looking carefully at your line relative to mad jew, one can see they're closer in 2016-07-03, and they diverge over time. Compare that divergence to Eidorian vs. mad jew, where they look nearly parallel. If mad jew weren't drawn with a dashed line, one could easily mistake both him and Eidorian as being low-rate or inactive users.
Now let's see the graphs as I omit high-count posters one by one, starting with maflynn:
Here, you're orange because colors are automatically assigned. Your line still looks fairly flat to me. The divergence from mad jew is still plain to see.
Now with C DM also omitted:
Here, you're cyan, but still look fairly flat, unless one looks for the divergence.
This graph illustrates that GGJStudios has had a fairly low post rate for a few years now, at least in relation to high-rate posters, or to overall post count. If one looks at Table 1, this is confirmed.
Here's the graph with GGJStudios also omitted:
Here's the first graph where you're not looking quite so flat. The line is still relatively straight (no bends), but with mad jew acting as an "artificial horizon" it's easier to see the upward slope.
And finally, here's the graph with all those above you omitted:
It's the least flat, but yours is still one of the more horizontalish lines.
If I were to use smaller groups, say the Top 10, Second 10, etc, your line would still look fairly flat, as illustrated by the first example graph (Top 7).
If I did use these smaller groups, the apparent flatness will depend on who you happen to be grouped with. Because of your post-count position relative to both some high-rate posters and some low-rate posters, I don't know of a simple or obvious way to group the graphable users that isn't confusing, or that won't need manual intervention.
For me, the fact that all the graphs and tables are produced automatically using a handful of scripts is important. There are a few places where I still have to manually make some judgements, but I'd rather not add more of those.
As an example of the importance of automation for me, I produced all the example graphs in this post just by editing a '#' into a few lines in one file, then running the entire graph-producing script (a one-line Terminal command). This generated almost 20 graphs in about 6 seconds, only one of which I wanted. The ease of editing the one text file, and the subsequent simplicity of moving the one graph I wanted to another folder made the overall task of graph production trivial. It took much longer to write the text of this post than to produce the graphs for it.
I'm open to ideas for ways to group users in ways that make nicer looking graphs. Just writing up this reply makes me think that an "Omit N" feature, where N is a number manually chosen to "look nice", might be worth exploring.
I would often pick one user and graph their entire history since joining MacRumors. Since I wasn't tracking their post count at first, until they had sufficient posts to appear on my radar, I'd assume that they posted at a steady rate from their join date to the date that I started tracking them. But the rest of the graph would reflect their actual post rate over 6-month time intervals, and you could see how it accelerated or decelerated.
Is that something you can automate too?
Is that something you can automate too?
Clever stuff. I’m glad it’s automated for you. I wouldn’t have a clue where to start.
Yes, in principle.
Right now it's mainly a question of available data. I only have a few years worth of data in a usable format. I still have the Enormous Spreadsheet you first sent me when I took over the stats, but I haven't tried pulling any more data from it. It's very unwieldy due to its size, and I've been reluctant to put time into reducing it. That was one of the reasons I stopped doing the One User Highlight graph, and started doing more frequent data (weekly captures) and graphing post-growth (prolificness).
Below is an example I threw together of a single user's detailed post-count history. It uses the scripts and data I have now.
First, I chose the single user to graph: you. I put that user's key info into the "keys file" that tells the scripts who to graph. Here's the single line in that keys file:
This is basically your post-count rank (7), user-number (12149), status (OK), name (Doctor Q), then some fillers and placeholders to satisfy the various scripts.
Next, I setup the "capture files" that have the post-count and other data in them, taken on past dates. The most numerous of these are the weekly captures dating back to early October 2018. It would work the same if I used past semi-annual dates, I'd just point it at a different folder of dated files.
Finally, I ran the script that reads the keys file, extracts user entries, puts them into a graphable form, and runs the actual graphing script that produces a PDF. That script ran in 0.585 secs (real) according to the 'time' command.
The PDF is then run through an automated conversion script built on the 'sips' command, which converts all the PDF graphs to GIF.
If I want the bi-weekly post count, all I need to do is change an environment variable from 1 (weekly) to 2 (bi-weekly) and run the main script again. About one second later I'd have this graph:
If I had to do this regularly, I'd set it up to take a command-line arg for the user-number, rather than needing me to manually edit a keys file.
EDIT
For comedic purposes, I put 3 entries into the keys file and ran the script.
It's Flatland:
First, I chose the single user to graph: you. I put that user's key info into the "keys file" that tells the scripts who to graph. Here's the single line in that keys file:
Code:
7 12149 OK Doctor Q __ __ 35442 257Next, I setup the "capture files" that have the post-count and other data in them, taken on past dates. The most numerous of these are the weekly captures dating back to early October 2018. It would work the same if I used past semi-annual dates, I'd just point it at a different folder of dated files.
Finally, I ran the script that reads the keys file, extracts user entries, puts them into a graphable form, and runs the actual graphing script that produces a PDF. That script ran in 0.585 secs (real) according to the 'time' command.
The PDF is then run through an automated conversion script built on the 'sips' command, which converts all the PDF graphs to GIF.
If I want the bi-weekly post count, all I need to do is change an environment variable from 1 (weekly) to 2 (bi-weekly) and run the main script again. About one second later I'd have this graph:
If I had to do this regularly, I'd set it up to take a command-line arg for the user-number, rather than needing me to manually edit a keys file.
EDIT
For comedic purposes, I put 3 entries into the keys file and ran the script.
It's Flatland:
Last edited:
You have the processes much more automated than I ever did. That makes it easier to slice the data various ways to look for interesting results. Thanks for the effort you put into it.
Absolutely fascinating series of graphs, and thanks, @chown33 for the time and trouble you have taken in compiling this.