Interpret these how you like. Half of the outliers either have typos or really jagged polars. These are at Re = 100,000, and include 1200 airfoils. The database copies very well into spread sheets. I only did thicknesses above 8% and L/D above 30 and camber above 1%, and I cut off some of the extremes so the graphs would be zoomed in. I checked out the extremes first, and it did not look like you'd be missing much. Most of the horizontal values are in % of cord. I'm not completely sure how they define trailing angle. Note: the Cl is Cl max, not Cl @ L/D max. You can't estimate the drag from the Cl and the L/D max, since the Cl is the Cl max, not the Cl @ L/D max. Cl is included only for stall speed. The strongest correlation here is that increasing camber increases the coefficient of lift. As for airfoil thickness and Cl, I think the only reason we see a correlation at all is that most of the thicker airfoils also have higher camber as a percentage of cord.