How to visualise multiple (different) sets of data in a same graph

Tags: #<Tag:0x00007f5ed361bc80>

I have cognitive data from Cambridge Brain Sciences (4 scores: c-score, reasoning, verbal and memory) which I gather every day, a couple of times per day.

I then have things like: hours slept that day, sleep phases, etc.

Next to that, I have supplements I take on a given day (or I don’t take).

I am wondering how could I combine all of this data in the same graph in a meaningful way. At the moment, I am tracking CBS data in a google spreadsheet with some basic graphs, but I don’t know how to also draw nr of pills (and type) and hours slept (and duration of each phase) all in the same graph for easy correlation.

If that’s not possible, how would I create the google graphs for the pills take for example? At the moment I have a table such as this:

Date Supplement Count Amount (per 1 unit)
2019-10-10 08:00:00 Magnesium 2 100mg
2019-10-10 08:00:00 Vitamin C 2 1000mg
2019-10-11 09:00:00 Omega 3 3 500 EPA, 250 DHA
2019-10-11 09:00:00 Magnesium 3 100mg

But I’m not sure what graph in google spreadsheets could display, on a per-day, per-hour detail, what did I take and how much. Any ideas?

To do a rough “correlation by eyeball”, you don’t need to squash the data into a single chart; just make sure all the charts line up on the x/time axis. Don’t know if that’s possible with Google Charts, but you can do it in Tableau, or if you are building your own charts with Python/Altair.

If you want to figure out if the supplements affect the cognitive scores, keep in mind is that the effects of some supplements may be cumulative. Also, to establish cause and effect, you might need to wait for your cognitive scores to reach a somewhat stable state, then add/remove supplements for some periods of time.

It sounds like you have a variety of data, @throw_my_username! That probably means there’s a lot of different ways you could imagine visualizing it. Another idea is to try things other than timelines.

Is there some specific question you have, or a pattern you would expect to see? e.g. if you plot “hours slept” on one axis and CBS scores on another – maybe you expect to see correlations there?