Today passed by so quickly. It was not until a few moments ago that I had the time to start writing. Right now Pandora internet radio is fully booted up and I’m listening to some music and beginning to sit down and write. That is good news at the end of a very long day. Sure I’m already thinking about dinner tomorrow, but that should not stop me from a little bit of deep thought and maybe a few moments of reflection. Some of those thoughts tonight need to be spent on the next iteration of my applied machine learning roadshow. Maybe working out the mechanics of how to run it from end to end is a worthwhile approach to explaining the full picture. Anyway — I’m going to spend a little bit more time digging into expanding on my initial 5,000 word treatment of the concept. Maybe take one good shot at it a day is the right way to move things forward on that front.
One of the things that I need to do as we approach 2020 is work out the three paper topics I want to start working on next year. Maybe really digging into modern approaches to survey techniques might be a place to start. I’m pretty sure that the modern methodology for gathering preferences is breaking down. You can always pay for focus groups, but random calling via telephone or sending out written survey cards is really becoming problematic. People are fatigued and generally not willing to be respondents. I hung up on a survey the other day for a new reason. I wondered if they were just collecting preference data to help them sell me things vs. doing some type of legitimate polling exercise. Pollsters are not registered by the state like lawyers or medical professionals.
Well that line of thought took off in a few directions. Most of them revolved around how to setup a new polling entity and what would be the general cost and consequence of doing that sort of thing. One of the things that I started to build out the other day and stopped was a real time sentiment extraction tool for sampling the news and social media. It seems like a ton of those types of tools have been built by stock traders and other groups of people looking to better understand the direction of the public mind in near real time. Even the idea of taking on that type of thing is a relatively interesting endeavor to consider.