This experiment explores diabetes data over time. This visualization slices glucose data into months, weeksn, and days.
In the example below, notice the two highlighed glucose values across month, week, and day time ranges. Search for these outlier values is facilitated by animating the transitions between time ranges. This animation allows the user to track specific markers even as the range of time shifts.
Looking at the day range, a pattern emerges. A high glucose reading in the morning leads to a day of borderline lows. Conservative insulin doses all day lead to another high in the evening. Using this search pattern, I've discovered this recurring trend of over-covering, under-covering, and subsequent high readings, starting the pattern all over again. I have modified my insulin doses on days that start out like this by avoiding over-conservative insulin doses after my glucose normalizes.
Dividing the data set in this way allows for insight into different metabolic rhythms and contexts. The week view brings to mind patterns related to meals, work, commuting, and sleep. Looking at a year's values brings light to seasons and gradual life shifts.