Amelia Greenhall on Using Moving Averages for Maintenance

We’ve posted some great talks by Amelia Greenhall here on the blog and we’re excited to bring you another insightful presentation. Last year Amelia gave a wonderful talk about her weight loss journey and the power of using running averages. In this updated talk Amelia gives a more in-depth look about how using a 10-day moving average serves as an “early warning system” that puts helps put her back on the path of mindful eating. Filmed at the QS Silicon Valley meetup group

This entry was posted in Videos and tagged , , , , , , , . Bookmark the permalink.

11 Responses to Amelia Greenhall on Using Moving Averages for Maintenance

  1. This is similar to what The Hacker’s Diet recommends. (Which is also what Beeminder does.)

  2. spg says:

    Fascinating! I have been using a 14 day rolling average to track around 24 different items including weight, body fat percent, blood glucose levels & blood pressure.

  3. One of the things I really liked here was Amelia’s comment that sometimes a line that doesn’t change is interesting and useful, too. What she means, in this case, I think, is that at the right scale and with the right transformation (in this case a simple moving average), what might have look like “noisy” or “random” or “static” data becomes coherent and meaningful.

  4. Tucker says:

    Very interesting discussion. While the use of any moving average period (10, 14) can be beneficial for a number of reasons, I personally believe the period of choice should have some significance in terms of ones cyclical behavior since the resulting calculation shows the typical value/trend for that period. This is why I use a 7-day simple moving average…so much of our lives are based around a weekly cycle and the resulting value tells me how my behavior has progressed over the last week. One could also argue that a monthly average is relevant in this regard, yet I’ve found that ~30-day averages introduce to much lag to be useful.

    • The Hacker’s Diet (where I got the idea from) makes a pretty conscious choice to use a 10 day average (see the Signal and Noise chapter in the section “Simple moving averages”, but also gets into exponentially smothing averages, which you could use to reduce the lag for a long period. I think the difference between 7 and 10 is mostly that 10 removes more of the jagged up and downs. Maybe I’ll recalculate my data later and see what the differences in the lines are.

      • Tucker says:

        Sorry, I missed the description of the reasoning behind the 10EMA, my 10 mo. old daughter must have pulled me away from my screen during that portion! A 10EMA is similar to a 7SMA in that it responds more to recent prices, making it less laggy that a 10SMA.

  5. Pingback: I’m a builder. | I Do Startup

  6. Pingback: In Search of a Universal Self-Tracking App

  7. Tim Chester says:

    Hi Amelia,

    Really interesting stuff. I’m writing a feature on QS for the Sunday Times in London and am looking for inspiring case studies. Would it be possible to email me on tpchester@gmail.com – I have 3 questions I’d like to ask if so.

    Thanks! Tim

Leave a Reply

Your email address will not be published. Required fields are marked *


*

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>

Notify me of followup comments via e-mail. You can also subscribe without commenting.