This density map is more revealing in the pattern of waterfalls on the landscape then the next map. This seems to reveal considerable sampling bias with major hotspots associated with public lands like the Big South Fork NRRA and the Great Smoky Mountains National Park. While this is to be expected, any dataset should be looking at itself closely and considering where its weaknesses are. Many folks contribute to the webpage Tennessee Landforms, myself included. I know my own personal bias tends towards the Upper Cumberland region where I live and work. Other waterfall hounds have their favored stomping grounds as well. Perhaps this should just be a reminder to occasionally get out of your comfort zone.
The data of this map is notably more difficult to interpret. I was attempting to see if spatial patterns for waterfall height were obvious. The first problem is figuring out how tall a waterfall is. By my measurement one may be 100' and by yours it may only be 30'. There are plenty of ways to see how this may lead to a dataset with inconsistent data. Further, there are waterfalls with no height data. This map captures mean height, the standard deviation of height (small dots show little variability, large dots a lot), and finally the number of waterfalls without height data, all by the same unit, a 500 square mile hexagon.