We are building an analysis program for one of our clients that looks through sales history over 3 years and graphs customer buying trends.

We'd like to include a straight best-fit trendline that is based on a "sum of least squares" formula. We are using ChartFX for the graphing. We actually need to calculate the formula for the trendline before we graph it. We want to filter the display to only customers with up or down trends (positive or negative slopes).

I know this is a complicated formula, but does anyone know how to calculate the slope and intercept (y=mx+b) for this trendline? Oh, and it's been way too many years since my calculus classes to do anything but keep it simple <g>.