More seriously -- and here comes the intellectual point -- is the variance between different map interpolations based on thresholds and assumptions. Using all cities where more than one user uploaded how much they paid for weed, we interpolated the following map (using kriging based the 12-nearest points). This demonstrates the highly nuanced topography of marijuana pricing but as some cities are only represented by a single report the reliability of this nuance is more suspect. (Note: we did eliminate all price reports that were two standard deviations above and below the national average.)
For example, the strip of low prices to the East of Syracuse in upstate New York is sparsely populated and contains very few price reports.
Cost surface based on cities with at least one price report, n=>2
(click on map for a larger image)
Using only the cities where 10 or more people uploaded price reports results in the interpolated map below (also using kriging based the 12-nearest points). This provides a more smooth surface for the variance in prices as price as there are fewer points and the prices between cities are estimated. In this map the low price region east of Syracuse, NY drops out....But interestingly other patterns in the previous interpolation such as the lower prices in central south Pennsylvania around Harrisburg remain. Moreover the lower prices west of the Appalachian mountains is consistent across both maps.
Cost surface based on cities with at least one price report, n=>10