List the measured values for each week. Label each measurement as "xi" where "i" is the number corresponding to the week which the measurement corresponds to. For example, if you are analyzing your money earned at the end of three weeks in a month, you will have three data points: x1, x2 and x3.
Label the number of weeks as "N." For example, calculating the variance over 3 weeks, would be "N = 3."
Calculate the mean average over the period of weeks. Sum all of the measurements and divide by the number of weeks. As a formula, calculate "Sigma(xi)/N," where "Sigma" refers to the sum over all of "i." Call this mean "m." For example, assume N = 3 and we have the data points 31, 41 and 59. The mean is then (31 + 41 + 59)/3 or 43.66.
Subtract the mean from all measurements. That is, create a new set of measurements, "yi." For each "xi," "yi = xi -- m." In our example, this step would yield the three points -12.66, -2.66 and 15.34.
Square all "yi" values. Multiply each "yi" by itself. Call these values "yi2." In our example, this would yield the three values 160.28, 7.08 and 235.32.
Sum the "yi2" values. Add every "yi2" together, resulting in a single number. Call this number "Numer." In our example, Numer = 160.28 + 7.08 + 235.32 = 402.68.
Compute the variance over the period of weeks. Divide "Numer" by the number of weeks, "N." The result is the variance you are seeking. Thus, for our example, the variance is 402.68/3 = 134.23.