Regression is a statistical method to evaulate the relationship between one variable and another. If two variables are correlated, as one changes value, the other will tend to change value as well. This allows us to spot trends and predict future values.

For example, in a typical financial application, we may measure sales versus last year to give an idea of whether our store is performing well. If sales versus last year is positive, then we are performing well. However we can evaluate this data further. Perhaps sales versus last year is positive, but trending downwards towards the end of the month. This suggests that next month's number may be negative. Regression analysis can spot this trend and provide a prediction for next month. Instead of looking at a single historical value, you can see that your sales are up 3%, trending downwards and next month is predicted at -1%.

In the sample graph above, the dots are the daily sales values. The red line is the calculated regression line. We can calculate a linear regression line or a polynomial, as shown above.