Thursday, January 26, 2017

Stat Badger

8,181,378 (as of Jan 26, 2017)

In 2015 after the last TAM, I was challenged by Jay Diamond to come up with a way that GSoW could show impact. Being that we work on Wikipedia pages, and readers are anonymous, how is it possible (unless someone specifically says that they changed their mind after reading a Wikipedia page) to know our results? 

We had one option, some talented non-GSoW code writers have come up with a way to "pull" numbers of pageviews. This is available to anyone, but you must have the exact spelling of the Wikipedia page when you search. You can compare several Wikipedia page views to other pages, you can select languages and dates. This is a quite useful tool. Play around with this, in fact I know you are going to want to bookmark this site. 

This is all great, but back to Jay Diamond's question, how can we monitor GSoW's results? Today is January 26, 2017 and according to our work spreadsheet we are well over 350 pages we can take credit for. And in several different languages. This would be complicated to use the Pageview Analysis tool to see results. Complicated to know when the page became a GSoW page.*

*(NOTE) We know that GSoW does not "own" a page. For the purposes of this exercise what we mean by "claim" is the day we made live a Wikipedia page that was created by GSoW or rewritten by GSoW. And we are not including pages we only work on, but pages that we have made extensive changes to, or written and released as a completed page. GSoW never releases stub pages, what we are talking about are pages that an editor has spent many hours fussing over, the page will probably include at least 15 citations, usually more. Once the Wikipedia page is made live, then of course anyone (you don't even need a Wikipedia account) can edit. Also keep in mind that GSoW has thousands of other edits, very important but we can't count every edit we make on Stat Badger. I had to make a choice, otherwise managing this would become a full-time job. 

So with that cleared up, I approached friend of GSoW, Kyle Polich who hosts the Data Skeptic Podcast and has also lectured at two of the Monterey County Skeptic SkeptiCamps. Keep in mind that I personally have no experience with building this kind of thing, and he was very patient with me asking for things I have no knowledge of. In about eight months we had a working version of Stat Badger. 

Now eighteen months later we are nearing the end of working out all the bugs. I've learned a lot about what is possible and what is not. In the last week or so Kyle was assisted by Mark Honeychurch who is the Chair of the New England Skeptics in getting the bugs out. What I'm looking at currently is something I'm so proud of. It was extremely tedious for me to decide what pages would make the cut, what day we could "claim" them as ours. Tag them (I'll explain in a bit) and input them into Stat Badger. Many many hours. 

I currently have pages tagged in categories which might be helpful when speaking to select audiences. Also it was interesting to see the themes of what interested my editors. If needed I will be able to delete tags or add more or whatever is needed in the future of GSoW.

One more thing I need to make clear. These are pageviews. That is all we know. We cannot know if the person accessing the page is a unique user or visits the page many times. We do not know how long the person has stayed on the page. We do not know if the page was visited by a bot (this was true for several years until Wikipedia changed its code) These are the restrictions that we are faced with. Stat Badger only can count how many times the Wikipedia page was accessed. 

Also I want readers to understand, GSoW editors normally work on pages they find interesting. These are not pages picked to increase stats. Many pages we work on receive very few hits, some under 100 a month. We are a volunteer organization, we have work lists but they are not necessarily ranked. We have discovered that a quiet Wikipedia page attracting very few views may suddenly sky-rocket because of something happening in the real world. A great example of this is when we rewrote the Nathan Phelps Wikipedia page and days later we were surprised to find that his father Fred Phelps died. 

For tags I currently have countries; Australia, New Zealand, Canada, Dutch, Hungarian, Russian, UK and USA. 

Languages; English, Dutch, French, German, Hungarian, Italian, Portuguese, Russian, Spanish and other.  I have people in training at the moment in many other languages, as they start working through pages, we will add more languages. 

"Our People", "Our Groups",  "Our Print" and "Our Audio" 
Astronomy, Cryptid, Medical, Paranormal, UFO, VAX, Book, Atheist

Also I have categories for "CSI Fellows", "TAM Speakers" and associated with ECSO

I have also tagged for a couple editors who want to keep track of their personal input.
At the beginning of this project I thought I would be releasing Stat Badger for public viewing. At this time I have decided not to release it. Only GSoW editors will have access to these numbers. I might change my mind in time, but for the moment we are going to keep this internal. If we decide to release this it will probably be hosted on the Data Skeptic website

So here are some results (this is not a fully accurate list as we have about 15 pages that for some reason are not pulling up numbers and all of the Russian pages are not yet pulling numbers) 

These numbers are from Thursday January 26th, 2017

Total pageviews 8,181,378

Rewritten pages 7,078,370
New Pages 788,785

Astronomy 636,817
Paranormal 2,377,693
Atheist 789,382
VAX 111,632
UFO 81,189
Cryptid 355,847

Dutch 435,842
Italian 52,000
Spanish 369,417
English 6,920,155
French 3,038
Portuguese 290,266
German 92,165
Hungarian 3,237 

Our People 4,635,086
USA 3,716,795
Dutch 291,955
UK 539,855
Australia 102,158
Canada 583,847
New Zealand 5,061

Our Groups 291,955
TAM Speaker 1,868,837
CSI Fellows 1,167,593

Did You Know? 609,951
Books 94,611
Our Print 164,387
Our Audio 123,047

ECSO 44,299
Medical 259,203

The top ten pages that have received the most views in the last month are... 

Spontaneous human combustion with 45,416
Jenna Miscavige Hill 37,967
Montack Project 25,021
Dan Ariely 15,619
Dean Cameron 13,463
Holocaust (Dutch) 12,873
Chupacabras (Spanish) 12,030
William B. Davis 11,307
One Million Dollar Paranormal Challenge 11,083
Arjen Lubach (Dutch) 10,845

I find this all very interesting. I love to browse over Stat Badger looking for oddities. I've heard many times that the skeptic community should not bother with pages for Bigfoot and Lake Monsters, "Everyone already knows they don't exist" But that isn't true, these numbers at least show that people are interested in the topic. The Chupacabra rewrite we did in Spanish in Feb 2015 has already been viewed 352,753 times. Spontaneous human combustion 1,058,399 views since we rewrote it August 2013. And the Montauk Project 490,031 times. Of course people still find the paranormal interesting. And we need to make sure that what they are reading is accurate with great citations. 

I'm hoping to announce numbers every so often, stay tuned

Susan Gerbic