User talk:LArron

From RationalWiki
Jump to: navigation, search

Archives: 1, 2, 3, 4

Yes, this is the right place to ask for a graph, an update or a statistic about Conservapedia, RationalWiki, etc.


Is there anything with which I could help out, like, with graphs and stars and RW:AU and shit? I mean, I can't guarantee I'll have the software, the time, or even the knowledge of how to do any of it, but still. THE GREAT RIGHTEØUS DESTRØYER Dancing on the corpses' ashesDolan.png 13:17, 3 September 2014 (UTC)

Thank you very much! Please, have a look at RationalWiki_talk:Active_users#June.2C_July_and_August.... --larron (talk) 06:11, 5 September 2014 (UTC)

So, like… what do I do? I don't understand. So, do I need some form of programme, or HTML skills, or…? MESSIAH OF DOOM Unite with thy oracleDolan.png 11:14, 5 September 2014 (UTC)
Hee! Google Extracting data from a wiki - good luck! (I haven't a clue myself and have nothing but admiration for those who do it) Scream!! (talk) 11:31, 5 September 2014 (UTC)

*prods body* Are you still alive?[edit]

In the remote chance of you having subscribed to changes to this page via email, there is an argument about user activity at the Saloon Bar, including attempts to quantify it. Your input will be appreciated. And I will appreciate if you bequeath to me the Perl routines used to populate RW:AU, or at least some pointers on how exactly you pulled the data (the idea is to have consistent results by using the same method).--ZooGuard (talk) 13:04, 1 February 2015 (UTC)

If LArron gets these messages and hands over his Perls, it'd be useful if we put them somewhere that all RationalWikians could access, so that it never happens again that nobody has the codes to update that page. oʇɐʇoԀʇɐϽʎzznℲ (talk/stalk) 14:22, 1 February 2015 (UTC)
If LArron turns out to be pining for the fjords, I will write something in Python and publish it as a subpage of the appropriate RW namespace pages. I just want to get the original algorithm right.--ZooGuard (talk) 14:32, 1 February 2015 (UTC)

Thanks for stopping by!

One of the reasons for not updating RW:AU is that I grew more and more unsatisfied with my way of gathering the data: I adopted the page in 2008 after User:Genghis_Khant had abandoned it, and back then, I used some scripts which I had mainly developed for Conservapedia, but used for other wikis, too. The approach is editor-based and works as follows:

  1. Look up current blocks
  2. Look up user groups
  3. Look up list of users
  4. For users whose edit numbers have changed since the last time, look up edit list

This approach has a great flaw when it comes to RationalWiki: it doesn't take into account anonymous edits! AFAIR, another RWian went via the list of pages, which includes anonymous accounts - a superior way. But this is unnecessary for RationalWiki, as we have the nifty active user list.

So, I mulled over redoing it, and it should be redone: now is the time! F*ck consistency, improve the procedure, and thank you, ZooGuard, for doing so!

And now my rant as an oldtimer (funny, I'm so old, I always thought of me as coming late to the party, joining RW only in 2008...):

I'm missing innovations. There was a time when new features popped up, sometimes obnoxious (liquid threads), but often eye-candy, while on other wikis, the time stood still. And now? We should have a mobile version. We should have an app: a simple version could just give you WIGO:CLOGS, WIGO:BLOGS, etc., and allow you to vote and keep you updated - it would instantly be better than Larry Sanger's infobitt!

At the moment, we are falling behind the times. Our voting pages are seven years old, they were fun at that time, but now they seem awfully static. And they have so much potential!

--larron (talk) 17:15, 1 February 2015 (UTC)

I need some open source development chops. That sounds like the kind of easy to develop application that anyone could do. Maybe making that is a challenge I should consider. Ikanreed (talk) 21:52, 5 February 2015 (UTC)
Great! --larron (talk) 15:41, 6 February 2015 (UTC)

R chops[edit]

You might've seen my disgusting packed circle graph for 2016 December. Here's the code that produced it. The current strategy was just to visit each diff (not even using the API, for shame) and pull data about the diff using tag classes. As a broad question: do you have any suggestions? Fuzzy "Cat" Potato, Jr. (talk/stalk) 14:11, 4 January 2017 (UTC)

  1. I'm gathering the data in a similar way: I monitor the revisions (not the diffs) and put the information in a database (together with information from the logs). I created the program originally for conservapedia - I didn't want to miss all these deleted revisions.
  2. here is my version for Dec 2016:
  3. using svg allows you to put much more information into the pic without it becoming to big: yours is 672kB, mine 282kB....
  4. I put the edits of all namespaces together for each title (e.g, Pizzagate and Talk:Pizzagate) to reduce the number of bubbles...
  5. the method I wrote to calculate the position of the bubbles allows for various subgroups, it puts each new bubble as near to the centre of gravity of its group as possible, thereby displaying the groups somewhat separately.
  6. simple improvement: have the bubbles ordered by size before applying your positioning algorithm....

--larron (talk) 14:47, 4 January 2017 (UTC)

Thanks, all of those are solid ways to improve. My graph is obviously deficient, I wasn't sure how to improve. As an additional question: how do you create each of the bubbles as moveable pie charts? Fuzzy "Cat" Potato, Jr. (talk/stalk) 15:06, 4 January 2017 (UTC)
I just saw that #6 won't work: you are not using a deterministic algorithm to place the bubbles...
Well, I like the often maligned pie-charts. But the R function "pie" is very insufficient, so I wrote my own (crude) version which allows for choosing center and radius:
Pie.variable <- function(x,y,r,div, angle=0, n=360, col="grey75",...){
	if(length(div)==0) return(FALSE)<br>
		H <- 2*pi*((1:n)-1)/n<br>
		X <- r*cos(H)+x<br>
		Y <- r*sin(H)+y<br>
	col <- rep(col,length(div))<br>
	H <- 2*pi*((1:n)-1)/n<br>
	X <- r*cos(H)<br>
	Y <- r*sin(H)<br>
	div <- c(0,div)<br>
	Div <- 2*pi*cumsum(div)/sum(div)<br>
	H <- 2*pi*((1:n)-1)/n<br>
	for(k in 1:(length(Div)-1)){<br>
		Auswahl <- H > Div[k] & H < Div[k+1]<br>
		PfadX <- c(0,r*cos(Div[k]),X[Auswahl],r*cos(Div[k+1]),0) + x<br>
		PfadY <- y + c(0,r*sin(Div[k]),Y[Auswahl],r*sin(Div[k+1]),0)<br>
		polygon(PfadX,PfadY, col=col[k],...)<br>

(sorry for the bad style, I use big letters and German names when I don't want to get things confused with the normal R code....) --larron (talk) 15:43, 4 January 2017 (UTC)

No apologies necessary! Thanks for sharing your code! :) I'm presuming that: x,y is the circle center; r is sqrt(editcount/pi); and div is the number of namespaces. If I might ask: how does the "col" variable work? Is it just the shadow?
I used packedcircles because of laziness -- all it does is ensure that circles aren't intersecting each other. I could conceivably coerce the package to fit your suggestions -- by providing certain initial circle values and weighting their movement -- but I'd be better off writing a deterministic function if I wanted to add your groups. oʇɐʇoԀʇɐϽʎzznℲ (talk/stalk) 22:12, 4 January 2017 (UTC)
col is the vector of colors used for the segments of the pie chart. It should be as long as div, but if it isn't, the colors will be reused.
The svg is just a text file. For this, I used
SVG.pie.variable <- function(x, y, r, div, angle=0, col="silver",...){
	if(length(div)==0) return(FALSE)
	cat('<circle cx="',x,'" cy="',y,'" r="',r,'" stroke="black" stroke-width="1" fill=',Col[1],' />\n',sep="")
	while(length(col) < length(div)) col <- c(col,col)
	Basis <- 0
	for(kk in 1:length(div)){ <- Basis+div[kk]/sum(div)
	Large.arc.flag <- 0
	Sweep.flag <- 0
	if(div[kk] > sum(div)/2){
		Large.arc.flag <- 1

	cat('<path d="M',x,',',y,
			'\nA',r,',',r,' 0 ',Large.arc.flag,',',Sweep.flag,' ',x+cos(2*pi*(*r,',',y-sin(2*pi*(*r,
			'\nZ" style="stroke:black; stroke-widht:1; fill:',col[kk],';" />\n',sep="")
		Basis <-

This produces just a line of text describing my pie chart. The shading is done separately, using

<filter id="dropShadow"><feGaussianBlur in="SourceAlpha" stdDeviation="3" /><feOffset dx="2" dy="4" /><feMerge><feMergeNode /><feMergeNode in="SourceGraphic" /></feMerge></filter>

on the group of pies. --larron (talk) 22:30, 4 January 2017 (UTC)

File:Ns-articles-RationalWiki-250-2016.svg was created like a png, but saved as a svg instead. The disadvantage is the size of the file, that was the reason to produce the svg as texts... --larron (talk) 00:04, 5 January 2017 (UTC)

DPL contributions on your user page[edit]

I'm going to remove the DPL contributions box from your user page, and I'd suggest that you leave it out at least until MySQL is upgraded. Under ordinary circumstances, it takes about 5 seconds, but when the server is under load, it is very slow indeed -- I logged your user page yesterday taking 16 hours (!) to render. The PHP timeout was not hit because it measures CPU time, not including MySQL query time. It's probably a MySQL bug, but there's not much we can do about that right now other than disable things that trigger it. Tim Starling (talk) 06:49, 8 June 2017 (UTC)

Well, thanks - I wasn't aware of this problem. --larron (talk) 20:58, 11 June 2017 (UTC)

Charts for 2017[edit]

For Conservapedia and RationalWiki, when you are free, please? Happy new year anyway. t 10:30, 4 January 2018 (UTC)

I don't think that there is enough interest to justify posting these charts: the first one got one comment, the second one perhaps six. --larron (talk) 20:27, 4 January 2018 (UTC)
The charts may not have gotten many comments, but many RW readers (myself included) think they are very interesting and enjoy seeing them. Please create the 2017 charts. Thank you. (talk)
Thanks for creating the RW chart. Would you please also create the CP chart? It would be nice to compare the two sites and see what CP is up to in the Trump Era. (talk) 02:18, 6 January 2018 (UTC)
done --larron (talk) 14:18, 6 January 2018 (UTC)
Thanks. I asked a question on the WIGO page. (talk) 02:49, 7 January 2018 (UTC)
Thanks a lot, LArron. I'm curious how these data were collected. Is there a special page on CP that conveniently lists the most edited pages, number of edits for each namespace, number of edits for each user (perhaps the CP equivalent of Special:Editcount), etc?t 09:40, 8 January 2018 (UTC)
Not that I'm aware of: I used a perl script to look up last year's revisions via the api - like

I filled a database and used it to gather the information for my R graphics.
--larron (talk) 11:03, 8 January 2018 (UTC)

Happy new year to you, LArron. Have a great one.t 14:15, 2 January 2019 (UTC)