What Are You Doing for Self-Quantification?
Right now, I am counting steps each day and I logged them on paper everyday, as close to midnight as much as possible. Only yesterday did I achieve my 10,000 steps goal. I have only been doing it since Sunday. This is my fourth day. Yesterday, I started logging my weight (on paper) to see if walking 10,000 steps will help me lose weight. Granted, it’s rather manual, but also easy to do. If I try to purchase a pedometer that syncs to your computer automatically, it will costs me 99 USD brand new. Adding wireless output to my weighing scale will cost money too. There are two conditions that could lead me to purchasing sophisticated solutions: I am loaded with money and or I am overwhelmed with data input.
For now, the important thing is that I keep doing it for 30-60 days for habit formations.
What are you doing for self-quantification?
Zeo sleep tracker
Every morning I write down on 2 index cards a number 1-5 about the previous day, mood/productivity and creativity.
Then I write down my weight on the scale.
a window-tracking daemon (arbtt) records computer time as spent on various windows
I do spaced repetition every day, which produces some numbers which could be useful for QS, enforced by Beeminder.
I’m currently doing a dual n-back/nicotine experiment, which produces another 5 numbers, also enforced by Beeminder
#1-3 is very easy, a few seconds at most when I go to bed & wake up. #4 is no effort to do after the initial setup, since it is run by a cron job. #5 is more work but is a years-old habit so really not a problem. I struggle with #6.
I lose my pedometer a while ago, but I plan to order another one in a few days, which I will add to #1-2/the index cards. Other than that, I think this is a sustainable routine and useful for future experiments.
Are Zeos worth the price relative to other similar things? I’ve started using Sleep Cycle, and while I don’t expect it to be anywhere near as accurate, I don’t know if the difference in accuracy is worth ~$250.
Apparently the current product most equivalent to the bedside unit I use is the Zeo Sleep Manager Pro which is ~$90. I don’t know how accurate it is compared to Sleep Cycle: I think some of the Zeo papers compare with standard sleep accelerometers but it seems like every cellphone accelerometer product (there are tons because it’s so easy to write) does things a bit differently.
What kind of pedometer did you use and how accurate do you think it was?
A cheap Radioshack one I found lying around the house; I don’t think it was very accurate, but I was more interested in trends and relative amounts than an absolute accuracy.
I don’t know what “a dual n-back/nicotine experiment” is. Could you expand?
http://www.gwern.net/Nootropics#experiment-1
See http://www.gwern.net/Nootropics#experiment-1 for the full writeup (I just finished analyzing it today).
I use Tagtime to ping me at random times for the purposes of gathering representative data about mood, productivity, energy, reaction time and time allocation. Every time Tagtime beeps or vibrates, I open and complete this Google Drive form (the process takes around 10 seconds). I don’t use Tagtime’s built-in tagging feature because it’s much less flexible and the collected data is much harder to analyze and manipulate.
I want to try this too. How do you access the form quickly from your phone? I have considered also giving position information. Also, in which time intervals do you sample? There is an obvious convenience/accuracy tradeoff.
Open the form from your browser and bookmark it. Then long-press on the home screen, click ‘shortcuts’, ‘bookmark’, then select the bookmark you just created. To make the form open even faster, install the Google Drive Android app and set it to open forms by default. Next time you click on the bookmark, it will open on Google Drive.
Tagtime doesn’t sample at regular time intervals; instead, it uses a Poisson process. I don’t know the value of the relevant parameters, but judging from past experience, the expected length of the sampling interval is around one hour, which seems to strike a good balance between convenience and accuracy.
So I have set up a form and a link on my home screen. How do you deal with when you sleep or otherwise miss several pings? Is there a convenient method of reading the ping times?
I was aware of it being a poisson process, I wanted to know the expectation value of the time between two pings, which you named as about 1h. Thank you.
You are welcome. I deal with missed pings by completing a separate form for each ping and doing my best to rate my state at the time of pinging. Reconstructing past episodes in this manner has been shown to significantly reduce recall bias.
I ignore pings during sleep, since my goal is to assess how I feel and perform while I’m awake.
Having now run the program almost two days I have a new question: How do you analyze the data? Do you do it in Drive or do you have your own little script to do that ready with confidence intervals?
I love self quantification!
I just do it in Drive or Excel, though I agree that a script might be helpful.
iPhone specific instructions: You can make the link to a form an icon like an app. Open the link in the web browser on the phone, click the middle button on the bottom, and choose ‘Add to home screen.’
It is an Android app. For the latter the process is similar.
Fixed some of your grammar. (Most notably, the “you’re” contraction is not used for most questions; questions usually reverse the word order to “are you” and so “you’re” doesn’t abbreviate it.)
Another Beeminder fan here. I log how much time I spent practicing programming.
As of yesterday, I’ve been putting 20 minute pomodoros and time spent meditating into Beeminder. I’m sorting out a system where each task on my todo-list is worth a certain number of points based not on how challenging the task is, but how difficult I find it to motivate myself to start.
I have thought about working on a similar system, too. Ideally, the points awarded for completing a task should be a function both of the intrinsic aversiveness of the task and of the benefit of actually completing it. One could then use Beeminder to commit oneself to scoring a certain number of points per day.
Some other people are already doing this. I had the vague idea for a while, but only bothered to actually do anything about it after reading Katja Grace’s post about Beeminder on OB. The relevant part is this:
I think getting too bogged-down in the details of the system could be harmful in the long run. So far I’ve got a TODO.txt with a table of tasks I need to do regularly, a table of tasks I need to do just once or very occasionally, and a table for each day where I list the tasks and tally the points. I’ll revise the points as I go, but the hard part for me is to get into the habit of using the list.
If you have an iPhone (I don’t), you might want to try EpicWin. (A friend just pointed me to this app.)
EDIT: There is actually a previous LessWrong thread on EpicWin and todo list “gamification”.
I track all of my “I should do this daily or almost daily” tasks/habits. (Review TODO lists, schedule things for tomorrow, chores)
How long I read for.
How many personal tasks I completed that aren’t considered chores.
How many commits I made on my code.
If I am studying something, how many hours I spent studying and the exercises completed (the latter bit isn’t of too much use, other than it being >0)
Which supplements did I take.
A rough sketch of what I ate (basically, the food items and rough macro-nutrient quantities)
If I exercised then I track almost every detail. When/What/Weight lifted/Sets.
I will also occasionally record what my “mental clarity” feels like on a scale from 1 − 5. I don’t yet do this often enough for it to be worthwhile, and I may just stop altogether unless I figure out a useful system.
The best thing I have found for my personal habit formation is https://chains.cc/ but there are a lot of “riffs” on this idea out there. I have one desktop on my computer dedicated to tracking all of the things important to me (chains, calendar, TODO lists, long-term goals). I have Google Calendar “event” that creates a popup on this desktop every day at 5:00am, so that ensures it is the first thing I look at.
Every day I measure my weight, systolic blood pressure, diastolic blood pressure, pulse, total sleep, ZQ (a point estimate of sleep quality), deep sleep, REM sleep, number of awakenings, the number of cards I review on Anki (along with how many I get correct/incorrect), quantity of melatonin taken, and quantity of vitamin D taken.
Every three months I measure my cholesterol and testosterone.
Starting on 09/01/2012, I plan to start measuring my type of workout, duration of workout, and quantity of nicotine taken.
I’m attempting to track my calibration over time by using PredictionBook. You can find the results here, but I stopped (at least until we get a decent tagging feature) making my personal productivity goals public due to Grognor’s complaints.
I have started analyzing some of my primary data on my personal blog here and am willing to share my raw data (just PM me if you want a copy).
I’m using rescuetime to track my work. I use pomodoro only if I can’t really get myself to start working on a task (helps only rarely), if I get working I turn it off because it brakes my flow… I keep a detailed todo list in org-mode.
I also keep a weekly gratitude list.
I track my weight when I work out and I’m near a scale, and I track my pulse rate when I remember to. I have a google docs form for each metric and a pretty little graph on each accompanying spreadsheet. I’ve been doing this for about 3 months.
Tagtime, mentioned in another comment, looks pretty cool and something I might look into at some point.
Daily I measure weight and workout performance.
Monthly I look at my receipts and spending and whether there were any large deviations from my budget.
Every six months I pick an unexceptional week and log everything I do by category (e.g. sleeping, preparing food, working, studying). I create summary tables for the typical week and use them to choose two or three improvements to implement in the next iteration.
I don’t do QS and I don’t know where to start.
Buy a pedometer and get some paper/index card and pen.
Make sure you put the paper and pen in a convenient location near your bed.
Set a goal, and wear your pedometer everywhere you go. Try to reach your daily goal. Once you go to bed, record your steps for the day. When you wake up, you repeat the whole process.
I’ve just returned to school and I’d not have time nor opportunity to go walking for an appreciable length time except on weekends.
One friend of mine simply sprinted between every class. Seemed to help his fitness!
also generates free time! generally just trying to walk between classes as fast as possible is probably good, if sprinting seems too scary.
Have a break between classes? Take a turn around the building. Tried this for a week, felt a little silly but I actually felt somewhat better.
Simplest and probably oldest example: Weight yourself, be it daily or weekly. I recommend the latter as it will better reflect actual changes in weight than a daily measurement in wich fluctuations through hydration dominate any real change.
Good point. Another method I’ve used is to average daily weights for a weekly average. If the average is trending downward over time, I’m losing weight. The other method is to measure only maximums & minimums. This method is psychologically reinforcing because you can have a little celebration (or a cheat meal, for instance) when you hit a new minimum (for those losing weight).
I use the Iphone Weighbot app to track my weight every day, with auto backing up to the website. That’s all I currently track.
I maxed out the 200 sit-ups app, and now I use the 100 pushups and 200 squats apps on alternate days and do 200 situps when I wake up and before bed.
I’m using beeminder to tally the number of 20 min Pomodoros I do per day
I have a 7 day calendar with all my repeating daily or weekly tasks on it (the situps etc, putting up a blog post, retagging 5 posts on my blog from before the migration that lost the tags, checking on Mint on Thursdays, attending Krav and aikido class, etc). When I do a task, it gets a dot, when I miss it, it gets an x. Days with no x’s get a happy face. On Monday, I count up the number of dots and log it in beeminder.
I’m 70, I lift three days a week, cycle 20 miles three days a week and eat less and have more sex. Read on Fox News last week! http://magazine.foxnews.com/food-wellness/want-live-longer-eat-less-have-mor