We've calculated the center of your digital suffering (mean and median) and identified your most frequent time sinks (mode). But knowing the average isn't enough for the algorithms that govern your life. They also need to know how predictable you are. This is where variability comes in – how spread out your data is. Are your screen times tightly clustered around the mean, or do they swing wildly from day to day?
Two basic measures of variability are the range and the variance. The range is simple: the difference between your highest and lowest screen time. It gives you a quick sense of the total spread. Variance is a bit more complex; it measures the average squared difference between each data point and the mean. A high variance means your data points are far from the mean, indicating less predictability. A low variance means they're clustered tightly around the mean – you're reliably miserable!
Algorithms thrive on low variability. If your behavior is predictable, they can optimize the content served to you, keeping you hooked. High variability, while perhaps a sign of a slightly less controlled existence, makes their job harder. Let's calculate the range and variance for our sample data and see just how predictable your digital habits are.
Observe the spread of the data points below, and see how the range and variance capture this variability.
The simplest measure of variability is the range. It is the difference between the maximum and minimum values in your dataset $X = \{x_1, x_2, \dots, x_n\}$.
Range $= \text{Maximum Value} - \text{Minimum Value}$
Using our example screen time data: $X = \{3, 5, 4, 6, 7, 5, 4\}$ hours.
Minimum Value = 3 hours
Maximum Value = 7 hours
Range $= 7 - 3 = 4$ hours.
In the visualization, the range is represented by the horizontal span covered by the data points.
The variance is a more robust measure of how spread out the data is around the mean ($\bar{x}$). It's calculated as the average of the squared differences from the mean.
Population Variance ($\sigma^2$) $= \frac{\sum_{i=1}^{n} (x_i - \bar{x})^2}{n}$
Sample Variance ($s^2$) $= \frac{\sum_{i=1}^{n} (x_i - \bar{x})^2}{n-1}$ (commonly used for samples)
Using our example data, with Mean $\bar{x} \approx 4.86$:
Calculate the squared differences: $(3-4.86)^2 \approx 3.46$, $(5-4.86)^2 \approx 0.02$, $(4-4.86)^2 \approx 0.74$, $(6-4.86)^2 \approx 1.30$, $(7-4.86)^2 \approx 4.58$, $(5-4.86)^2 \approx 0.02$, $(4-4.86)^2 \approx 0.74$.
Sum of squared differences $\approx 3.46 + 0.02 + 0.74 + 1.30 + 4.58 + 0.02 + 0.74 \approx 10.86$.
Using Sample Variance ($n=7$): $s^2 = \frac{10.86}{7-1} = \frac{10.86}{6} \approx 1.81$.
The visualization might show the data points' distance from the mean, conceptually representing the components of the variance calculation.
Variance is in squared units (hours squared in our example), which isn't intuitive. The standard deviation is the square root of the variance, bringing the measure back into the original units.
Population Standard Deviation ($\sigma$) $= \sqrt{\sigma^2}$
Sample Standard Deviation ($s$) $= \sqrt{s^2}$
Using our example Sample Variance $s^2 \approx 1.81$:
Standard Deviation $s = \sqrt{1.81} \approx 1.35$ hours.
A higher standard deviation means your daily screen time varies more from the average. A lower standard deviation means it's more consistent. The algorithms prefer a low standard deviation – it means your digital habits are reliably predictable, making you a more valuable (and controllable) data point.