2 dimensions (day of week + hour of day) + 1 metric
Allows us to detect a metric’s temporal evolution patterns on two different temporal axes: (a) during the week, reading the punch card top to bottom to see if a day of the week stands out for over/under performance, and (b) during the day, reading the punch card left to right to see if certain time slots are better or worse than others.
A simple example
Let’s say you want to know when people generally visit your website. For instance, you want to put an ad on your website and you need to find the optimal time when it will gain the most coverage.
Try it with your own data on the Visualization Playground
Let's analyze a bit
OWe can immediately see that people don’t visit your website at night or on the weekends.
We can see that each weekday, your activity grows starting at 9 am, decreases at around 1 pm, then peaks again around 3 pm.
Anyway, "Your Visits go through the roof on tuesday at 4pm"!