TagTime explanation vid

• Stochastic time tracking

• What is stochastic time tracking
  • Normal time tracking graph
  • Actual graph
  • Sampling from actual graph
• What is TagTime
• What is TagTime Web
• Time hashing

• {DONE}} "stochastic time tracking" for beginning/thumbnail
• {DONE}} Pros/cons table

• TagTime Web (my instance): https://ttw.smitop.com
  TagTime: https://github.com/tagtime/TagTime
  TagTime blog post: http://messymatters.com/tagtime/
  Graphs: https://alexschell.shinyapps.io/tagtime-vis/
  

word-count

• Stochastic time tracking is a new type of time tracking. Let's take a look at how it builds upon traditional time tracking systems. These systems usually involve recording what you're doing over constant intervals of time, usually each 6, 30, or 60 minutes. However, this system has some drawbacks. Firstly, most people don't actually spend their time in discrete chunks. Focus is shifted back and forth between things, quick things are done for just a few minutes, and things are done that can't really be categorised into per-determined categories. However, tracking what you are doing on a per-second or per-millisecond basis would be almost impossible, and would certainly result in more time spent tracking time than actually doing things. Generally, the more complex your time-tracking system, the more time you spend tracking your time (and then you'd need to track that time as well).
• One alternative approach to manual time-tracking is to automatically detect what you're doing, and annotate your time based on that. Services such as RescueTime do this by keeping track of what websites and apps you use, and categorising your usage into categories like Productive and Distracted automatically. This is a considerable improvement. However, there are some limitations here too. There's no generally-applicable way to decide if a given website/app is productive or distracting, since different people use them for different purposes. As such, there's a degree of manual classification required. Classifications can change over time as websites/apps are used for different purposes. Furthermore, what people are doing on a certain website is also relevant: professional networking on Twitter is Productive, but using it to read the latest news is Distracting (unless you are a journalist, in which case it is Productive). It's hard to capture the subtlety automatically. And it still suffers from overly broad categorisation. Still, automatic time tracking can be quite effective, and can certainly be an improvement over manual tracking. I'd recommend checking it out, and seeing if it works for you.
• Stochastic time tracking systems are an alternative method of time tracking, which is both efficient (as you don't need to spend a lot of time tracking) and accurate (with detailed tags, I'll explain that in a bit). It sends a ping at random intervals to you, asking what you are doing right at that moment. Then, instead of trying to fit what you are doing into a single category, you can enter multiple tags to describe what you were doing. Tagging is a much more extensible system than these rigid categories most other time tracking systems enforce. Most stochastic time tracking systems use an average ping gap of 45 minutes. This means that on average, each ping will be 45 minutes apart. This means every 24 hours, you get around 32 pings.
• Note that with stochastic tracking, pings aren't at constant intervals. Otherwise, it would be trivial to game the system by only working right before a ping occurs. It would also result in productivity going down right after a ping, since there wouldn't be another one for a while, so there would be no chance of that time being sampled. The size of the gaps between pings can vary quite considerably.
• While the data for a single day is worse than traditional interval tracking, there are many benefits. It's a lot harder to make mistakes in stochastic tracking systems, since you don't need to be constantly updating a system whenever you switch tasks. The time tracking system actively asks you what you are doing. But more importantly, you get really high-precision data. The time tracking system can accurately know how often you gaze off, or check your email. It gives you insight into if your email checking is a large distraction or not. There's also a lot of cool graphs that you can generate from a TagTime log. I have a link to one such graph generator in the description.
• The first implementation of such a system was called TagTime. It is an open-source implementation written in Perl. You might want to read the blog post about it from it's author (link in description). It's fairly minimal, but works well. I've recently written an open-source web version, called TagTime Web. I'm currently working on making TagTime Web work better. I have links to both in the description. I hope you found this interesting, thanks for watching.