April 2015 in review

This is the seventh of my month-in-review posts. I started month-in-review posts in October 2014. You can see my first post here.

Some overall background

As expected, this month was extremely busy at my job, so I ended up doing less of other things than I had hoped for (but within the range of my expectations). The coming month (May) will probably be substantially more busy, and I will probably not get time for either Wikipedia editing or writing blog posts or articles. Also, due to a relocation of our company office, I need to spend an extra 20-30 minutes walking to work every day, which will further cut into time I might spend on content creation activities.

Wikipedia editing

For background information, see my site page about Wikipedia

Wikipedia editing was high in the first half of the month, but relatively low in the second half. Much of the editing happened during BART rides, but I also spent a decent fraction of the first two weekends of April on it. The pages I created were largely in my focus areas of migration and effective altruism. Here are the pages I created (you can cross-check here):

  • H-1B1 visa, created Friday, April 3, 2015. It took 1-2 hours to create. It was also my first page about a United States visa.
  • Automatic visa revalidation, created Saturday, April 4, 2015. The page took about 30 minutes to create. It was relatively quick and painless but also correspondingly low-value.
  • Form I-129, created Sunday, April 5, 2015. The page took about two hours.
  • Premium Processing Service, created Sunday, April 5, 2015. The page took about 1-2 hours to create. It again arose out of my interest in better understanding the process and bureaucracy surrounding immigration.
  • The Most Good You Can Do, created Saturday, April 11, 2015. It took 36 minutes to create (clock time from start to end). Despite the short time to create, the page has been getting a decent amount of traffic. However, traffic is likely to decline over time. The page is hot right now because the book was released recently, on Wednesday April 7.
  • 10 Hours of Walking in NYC as a Woman, created Wednesday, April 14, 2015. The page was created as part of my effort to understand content that goes viral on social media as elsewhere, a topic of interest to me both professionally and personally.
  • The unique connection, created Friday April 24, 2015. As with the preceding page, this page too was created as part of my effort to understand viral content on the Internet.
  • Facebook Safety Check, created Monday, April 27, 2015. The page was created due to the topicality of Facebook Safety Check in the wake of the 2015 Nepal earthquake. It took about 1.5 hours to create, but much of the time was offline on BART (I was offline on my laptop, and had Internet on my cellphone, and was manually copying URLs and titles, which took longer).

I also made updates to these timeline pages to reflect recent events and changes:

I made significant content changes to these pages (not created by me):

  • H-1B visa: The edits here were to incorporate new information I had gathered as a result of creating many other pages related to US migration.
  • Effective altruism: I noticed that the pageviews for this page had risen quite a bit in late March and early April, likely due to the release of Peter Singer’s book The Most Good You Can Do. So I decided to finish a long-standing revamp of the Wikipedia page about effective altruism.
  • If London Were Syria, a viral video titled on YouTube The Most Shocking Second a Day Video.

One of the pages I was planning to create, Woven Digital, got published in main space just a few hours before I started working on my version of it. I was able to port some of the links I had collected to expand the existing page.

April impact: To better understand the impact that pages I had created or edited at particular points in time had at other points in time, I created separate tags for pages I created in each month. So now, it’s possible to access information on pageviews in a particular month (say April 2015) for pages I created in another month (say January 2015). Summing up along one dimension gives total pageviews by month of creation, whereas summing up along the other dimension gives total pageviews by the month of views.

A small caveat before I get into the April numbers. I discovered that a few of the pages I created, specifically SunSaluter, Tullock paradox, and Marginal Revolution University, had not been tagged in Wikipedia Views as pages created by me. Adding in these pages has therefore caused the pageview counts to change for previous months. The change is not significant for most months, but as it turns out, one of these pages (Tullock paradox) got an unexpected surge of traffic in April (5356 views, compared to 365 in March).

Page category Pageviews in April 2015 Pageviews in March 2015 Cumulative pageviews in 2015
Pages I created in April 2015 2473 101 2773
Pages I created in March 2015 2818 1752 4580
Pages I created in 2015 13134 7709 26136
Pages I created over my lifetime 177084 142994 573979

My forecast for the view count in April for all pages I created (bottom row, second column) had been 111,000-167,000 pageviews (80% probability) and 87,000-190,000 pageviews (95% probability). The actual value was 177,084. The count stayed within the 95% interval but was outside the 80% interval. The chief reason for the overshoot was that a few pages got a lot more traffic than expected:

  • Internet.org got 35,004 views, compared to 17,800 views in March 2015. The increase in pageviews was due to its topicality, starting around April 15.
  • Tullock paradox got 5356 views, compared to 365 in March 2015. There appears to have been an unusually large number of pageviews on April 22, perhaps because the page got linked from a high-traffic page.
  • Secret (app) got 5858 views, compared to 4425 in March 2015. The increased views were largely due to the app’s shutdown on April 29 generating news interest in it.
  • Pages I created in April got 2473 views, and pages I had created in late March saw an increase of 1066 in their pageview count from March to April.

May forecast: My point estimate for May pageviews is 151,000 (note that May has 31 days, compared to April which had 30). I estimate an 80% probability that the number of pageviews will fall between 120,000 and 200,000, and a 95% probability that it will fall between 100,000 and 250,000. The reason I’m choosing a lower point estimate for May than for April is that I believe April was unusual largely on account of Internet.org and the Tullock paradox, and previous months have seen lower traffic levels. At the same time, I’m choosing a higher point estimate than the number of pageviews seen in March. I have widened my intervals, and made them more asymmetric (skewed to be larger on the upside) because there is more potential for particular pages to get a lot more traffic than expected (due to topicality) than potential in the other direction. (I also computed view counts for May 1, which came to 5043, and would predict about 156,000 pageviews in May, but I believe this is a slight overestimate because some of the pages I created are still a little more topical right now than they will be for the rest of the month).

Subject wikis

General background information: My site page about the subject wikis

I spent zero time editing the subject wikis this month. Traffic to them grew at its usual steady pace, highlighting that the subject wikis have become sufficiently mature that they don’t need much ongoing supervision.

April impact: I’ve put the numbers for Groupprops in a table for easy perusal and comparison. Numbers are from Google Analytics.

Metric Apr 2015 Mar 2015 Feb 2015 Apr 2014 Mar 2014 Feb 2014
Pageviews 88,046 82,377 73,713 66,233 57,724 57,925
Sessions 43,968 40,280 36,417 34,310 29,249 29,123
Pages/session 2.00 2.05 2.02 1.93 1.97 1.99
Pageviews in sessions with at least 5 pageviews 27,991 27,641 24,338 19,837 17,924 18,354

The total number of pageviews (88,046) in April 2015 was safely within my 80% probability forecast interval of 74,000-103,000, and close to the point estimate of 85,000.

  • Market: 22,432 compared to 23,130 for the same month last year.
  • Calculus: 19,210 compared to 17,555 for the same month last year.

May forecast: Last May, Groupprops had 63,211 pageviews. Given the current year-on-year growth trend of 15-40%, my point estimate for May is 82,000 pageviews. I estimate that with 80% probability, the number of pageviews will fall between 72,000 and 96,000, and that with 95% probability, the number of pageviews will fall between 62,000 and 114,000.

Open borders

General background information: My site page about open borders

No blog post of mine was published, nor did I get time to work on any drafts. However, some of my Wikipedia edits related to migration were indirect preparation for future blog posts. I also did not execute on planned work for the site revamp.

You can see the site’s month-in-review post here.

Other sites and forms of online presence

  • This website (vipulnaik.com) got 2,161 views in the month of April according to Google Analytics (and 2,000 views according to WordPress Analytics), tying with January 2014 for the highest ever.
  • Wikipedia Views got 951 views.
  • Quora: For April, my views were: 13.9K on answers, 12.2K on questions, and 228 on posts. The grand total was 26.4K.

Entertainment

I bought two MP3 songs and watched the movie Ex Machina as part of a team outing.

Server downtime

There was no or negligible server downtime. The servers were up for 99.9%+ of the time. This was a welcome contrast to March, when the servers were down for what amounted to almost an entire day.