On Daily Mood Tracking

The only reason I’m writing this is that I let slip to a friend¹ that I’ve been tracking my mood daily for the past 4 and a half years, and I couldn’t credibly argue that the practice hasn’t conveyed some kind of expert status despite the fact that all I did was tell my phone how my day was whenever it asked me.

Most of these ramblings have only come to me within the past month as I’ve tried to corral my experiences into thought. They boil down to this: It’s not good to optimize for daily mood, both because feeling good isn’t the most important thing, and, even if you think it is, greedily maximizing it² isn’t the best algorithm to improve it, and even if you’re bound and determined to do it anyway, mood tracking probably isn’t the best way.

That said, I’m going to keep doing it at least until my current phone dies which could be any day now.

I used to not have bad moods. I’m sure I did when I was little, but by high school, things that bothered my friends, like bad grades, losing races, and traffic, didn’t get to me as much. Sometimes I did feel weird but not in a negative way. And I’m sure that was normal teenage weirdness. (Or else it was what happened when I missed my strict, self-enforced 9PM bedtime to finish a paper I’d been procrastinating).

This was a fact of life, regular as the seasons. I noticed it, thought it was odd, and described it to my high school girlfriend, but it never occurred to me to understand it until I saw a video³ about personal data tracking by an environmentally anxious Canadian who actually found a correlation between his mood and the weather.

I didn’t start doing it myself until early 2020. Maybe after a rough breakup, stalling undergrad research, and a cycle of injuries and PT from track, I finally thought I should figure out why I was feeling so bad. Maybe I was just going through a productivity guru phase (I also started using RescueTime⁴ around then). On February 11, 2020, I downloaded Daylio and recorded - Good (4/5) “Got up early to get groceries. Good long jump practice”.

Since then, every night at 11PM, my phone asks me how I’m feeling. I rate the day on a 1-5 scale (really “awful, bad, meh, good, rad” (with customizable colors and emojis)) and write a few words about my day. Since February 2020, I’ve gotten more verbose (or at least more active). I write about 20 words per day. There are checkboxes for every little thing you might have done (e.g. party, good sleep, fast food, drink water, kindness, laundry, and relax), but I don’t use them. They’re too much trouble for something I do every single day, and I don’t pay for Premium, which gives you the analytics about which checkboxes affect your mood the most.

As it is, I only use the free text field which exposes my little self experiment to an obvious bias. Since I choose what to include, mood tracking will only ever have data on events that I think are notable. Caffeine has a large positive impact on my mood, but I’d never mention drinking coffee except in passing if I had it on a date or with a friend. When people run actual studies like this, they send notifications at random times of day, asking what the person is doing and how they feel in that moment to get a variety of activities and moods.

That said, I do try to take it seriously otherwise, meaning I really do consult the way I’m feeling and try to record it honestly. I do the rating first before thinking about what happened and how I think it should make me feel. My mood is still mostly “good” with a smattering of “meh” and “rad” (there’s no “weird” option).

Unfortunately these biases stymie most analysis. If I paid for Premium, they would analyze it for me somehow. I’ve also considered running a dubious, homemade sentiment analysis. Fortunately, the actual, unscientific method I use is far more robust to these biases (because it comes with its own). I use the old school method of counting on myself to remember when I’m recording something I’ve done before. The biases don’t bother me since I haven’t put any actual work in besides introspection and reflection.

These are the things I’ve noticed that affect my mood since I started tracking it. Take them with a grain of salt.

My worst moods, which were most apparent during a couple project classes in school, come when I have work that I don’t like, care about, or know how to do that I need to discuss with a mentor or group. I’m ok with being stressed about a deadline, but there’s a specific kind of dread from disappointing people.

On the flip side, getting things done is the most consistent source of high mood⁵. It’s a positive feedback loop. In a low mood, I don’t feel like doing anything, so accomplishing something is evidence of a high mood which makes me feel like doing another thing. On top of that, the things I’m accomplishing usually matter to me, and I’m glad to complete them per se. For example, if I clean my apartment, I feel better because it’s nice to have a clean apartment, but I also feel better because I accomplished a task, and on top of that, it’s evidence that I’m in a good mood.

Better than a clean apartment is a good conversation with a friend. This is the opposite of dreading a conversation I expect to go poorly. I’ll include all kinds of serendipitous encounters in this category too. To be clear, I already expect conversations and walks to be good, but they’re that much more delightful when I happen upon a concert in the park or people water skiing.

Daylio is good at picking up on these flow and connections-type activities. I always mention how work is going and which friends I spent time with, so even my terrible analysis method has picked up on them through sheer quantity of data. But most people probably don’t need to put in all that effort to figure this stuff out. Overall, it’s not clear to me that mood tracking has any advantage over experience and introspection. At best, it’s just a forcing factor if you’re not naturally inclined to this kind of reflection.

Mood tracking may be worse than introspection. Tracking might obscure things that affect your mood more than it illuminates them. It’s easy to miss mundane things like diet, sleep, and exercise that have a large effect.

On the other end of the spectrum, there are moments of awe and inspiration that may be too rare to come through in the data. But those moments are so visceral that you would never need to write them down to know they improve your mood. The only reason would be to help you remember to pursue awe.

The highest moods I’ve ever experienced feel like overwhelming gratitude. Usually I feel this towards friends, family, and all the wonderful people in my life, but the first time was when I was leaving a bad day at work, merging onto the beltline, probably listening to Silk Chiffon⁶. It manifested as something like gratitude for being in a bad situation that I had the chance to improve. These didn’t really translate well to the app because they’re ineffable.

Even if you do find the things that correlate with higher moods, that doesn’t translate to reliably improving your mood. Mood tracking suffers from the classic statistical problem of not differentiating correlation from causation. If you know you dance on days when your mood is better, and you’re currently feeling bad, not being able to muster the energy to dance could make you feel worse.

The reason I don’t dance when I’m in a bad mood is not that I don’t know I enjoy dancing; it’s that I just don’t feel like it. I’d much rather sulk, eat ice cream, and watch reality TV despite knowing they’re correlated with low mood. When I’m in a bad mood, I usually let myself stay in it until some prior obligation forces me out (which is also why I’m opposed to canceling plans based on how I feel). If I’m really trying to actively trick myself into a high mood, I go for a walk (which I enjoy doing in a bad mood and elevates my mood and also never gets tracked because I do it every day).

Mood tracking is also blind to larger aspects of life that are constants in the long term. If you’re in a bad job or relationship, you might notice a consistently low mood, but you probably won’t mention it in your day-to-day tracking. Then, if you make any lifestyle changes based on your findings, that will always be missed. Hopefully you’re aware of anything big in your life that needs changing because trying to get to the Good Life by mood tracking is a classic case of trying to get to the Moon with taller and taller ladders.

To be honest, I’ve never tried to use it to improve my own mood. I’ve heard enough philosophy thought experiments to be wary of a drug that makes you feel good all the time, even if that drug turns out to be dancing. I can’t remember who phrased the age-old wisdom this way, but I subscribe to the idea that “the pursuit of positive experience is inherently a negative experience.” It’s a trap that has caught anyone who wonders if they’re currently doing the thing that is making them the most happy. Even if they were before, now they’re not. It seems like an obvious risk of tracking your mood down to 15 minute increments (Are We Happy Yet? by Jessica Grose for the NYT).

My saving grace was that I’ve just been curious since the beginning. It’s almost hard to imagine doing something just because I know that it’s correlated with a higher mood. Not to brag, but I have a pretty robust, built-in intuition for what I want to do, and I’ll still eat 3 Costco muffins for breakfast even if I logically know that it will make me feel bad and lethargic.

I also escape that fate by totally forgetting that I’ve been doing this every day for years. If I’m awake at 11PM, I still get excited that someone’s texted me before I realize it’s just my phone asking me how I feel. I never think about what decision will let me record a better mood for that day because I don’t remember that I’m doing it.

If I have made any lifestyle changes, I may be more active. I have no qualms with taking it easy for a day, but it gets embarrassing telling my phone I’ve been lazing around for multiple days in a row.

One of the most concrete positive impacts is that Daylio sends me old entries from previous years every Thursday morning at 11:24. It’s such a random time that it took me a while to realize it was the same every week. It brightens my day to remember good days, realize how far I’ve come, or get nostalgic. I’m a sucker for nostalgia⁷. Also, the notification itself calls them “New Memories” which is the funniest thing you can call old memories.

It reminds me of the time I discovered that Google Maps had been tracking and saving my location. It was so fun to look through and amazing how much I could remember about a day from just a little reminder about where I went. I was sad that I ultimately had to delete the data and disable the setting to protect my own privacy. Daylio is like that except it without Premium it doesn’t even automatically backup my data. Even if they did sell it, I don’t mind people knowing that I like calling my friends.