Some upcoming changes in daily stats …

Recently, I achieved what seemed to me to be a major milestone in my stats: I had 24/7 coverage of the major cable news networks’ shows. Every hour, even the crappy rerun of Tech Now in the middle of the night on Al Jazeera, was monitored, counted, and reported.

To do that, I have accumulated a huge database of shows and schedules.  There are 142  “active” shows in my database that I track, individually, and over 700 entries that map these shows to specific air times.  Last month (2/15), I processed over 4 million individual tweets, assigning each to one or more shows (you’d be surprised how many people will throw in hashtags for several shows in a single tweet).

I’m incredibly pleased with how comprehensive the statistics have become.  To achieve this, I’ve built some powerful tools.  I have a server that runs 24/7 with custom Java software to collect all the tweets in near real time and save them into a MySQL database.   The tweets are tagged by language, mood, and if discernable, the gender of the sender. Once a day I run another custom Java application that processes, analyzes, and generates the bulk of the daily reports I post.  I add in some commentary, build some charts when warranted, and a blog post is done.

If I could automate this 100%, I would probably run it indefinitely, but there’s still a huge, time consuming manual step in all this: cable news channels are constantly changing their schedules.  About 5-10% of the schedule changes every week.  CNN swaps around its documentaries.  Fox News has special reports, and Hannity tends to get shoved around the margins of the schedule.  Al Jazeera has so many half-hour shows that it’s more work keeping up with their schedule than all the rest of the cable news networks combined! Talent comes and goes from shows.

Every week I have to invest hours in reviewing the upcoming schedule, reflecting changes in my database, and keeping on top of new shows or changes in shows. That investment, I have concluded, is no longer worth making. There’s several reasons for that.

First, there’s no profit in this: I’ve tried putting advertising on the site, but it generates insignificant revenue.  I have contemplated some sort of subscription model, or tip bucket, but concluded nothing will produce enough cash-flow.  That tells me there’s little value in the work I do here. Or I’m inept in monetizing it, I suppose.

Second, the world of Social Media has fragmented.  Twitter is at a cross-roads.  Its growth has stalled while a myriad of competitors have drawn away users.  Reporting just on Twitter gives a fragmentary and, I suspect, inaccurate picture of the world. As an example, in the 2010 election cycle, Twitter activity correlated well to election results.  By 2014, that correlation had started to fall apart. Whether it’s because people are gaming Twitter, or trolling, or just the waning influence of the platform I do not now.

Third, by going “wide” — trying to be comprehensive — it keeps me from going deep into topics.  I would rather produce a couple of insightful and actionable blog posts a week than keep after schedule changes so I can produce daily reports.

Finally, I have other projects I want to work on that are more deserving of my limited time. I have some iOS apps that need to be refreshed, and others that I need to finish.  I have some data science projects that have been stalled.

So, I will be gradually phasing out the daily stats.  After Sunday’s numbers, I am immediately dropping HLN, Al Jazeera, Comedy Central, and HBO from the calculations. They’re interesting, but are a side show.  Then, I’ll reduce the amount of shows I track to just the very top tier of shows that get actual sustained engagement from viewers.  At some point, I may feel like the effort involved is trivial and stabilize there, but that point could well be no daily reports.

We’ll see.

If there’s something that’s actually important to you for, drop me a line at — I’ll see what I can do.