“Twitter can no more produce analysis than a monkey can type out a work of Shakespeare.”
Techcrunch wrote a small story on Eric Schmidt’s comment about Twitter. Must be a slow news day again 🙂
You know what, I think Eric’s right too. One comment I also agree with was a comment from David Finebaum
He’s simply referring to the signal-to-noise ratio on services like Twitter, which is aptly pointed out in that vodka ad. And he’s right. Twitter is free association at it’s core, a series of unrelated and irrelevant thoughts that are redundant in light of other services. Trying to draw anything meaningful form the ‘Twittersphere’ is a pointless exercise. I hear Nielsen is trying to gauge program viewership via Twitter now. I wonder what consulting agency sold them that snake oil…
And thus the onslaught of agree/disagree comments continues. Twitter has changed dramatically over the last five years since I started using it. My first though in September 2008 was, “why use this?”.
As a broadcast medium it’s immediate and perfect. As a discussion medium it’s next to useless.
Where’s Wally?
The reason I packed it all up is simple, distraction. I can’t afford to be distracted in what I’m doing. The orders of magnitude that are created every time our brains transfer from the task in hand to checking the tweets is too great. I always called Twitter my “office banter” as I essentially work alone. Now I’m in a team connected remotely but it’s more focused and concentrated on the task in hand.
I look at it this way. For every three seconds I glanced at Twitter I was killing 4 minutes of productivity. So for every three second glance I was losing four minutes for every 15 minutes worked (I might reply to a mention or a DM). That’s a potential 128 minutes for every working day. So for every working year (48 weeks) I’ll be spending 21 days and 8 hours on Twitter.
So I quit. Killed the lot….
As soon as I killed the signal to noise then everything changed. My productivity shot up dramatically. I know it’s easy for folk to say, “well just close it down” but it’s too tempting to fire up a twitter client and just have a quick check. In reality it changes nothing.
So how do I measure the broadcast?
I’m not overly bothered about the conversation, I can do that with email if needs be. For me email is still core of everything. So with a mixture of R, the twitteR library and sendmailR I pull hashtags that are of interest to me #hadoop, #retail, #analytics and get summaries at the end of the day.
Okay, it’s a rather low level way of seeing things but it works for me. There’s also blogs and RSS feeds that watch as well, though they suffer from the same problem – signal to noise. Out of 2000+ stories a day there might be only 10 I really want to read. That though is another post for another time.
It's All A Bet - MLOps, DevOps, AI and Synthetic Data Design 