California:
High Quality: $359.19 / oz. (n=72)
Medium Quality: $300 / oz. (n=17)
Low Quality: $648 / oz. (n=2) <-- ???
The low quality stuff is worth more than twice as much as the medium? The crazy numbers for low quality are on account of one wonky data point ($976/oz). There were also some obviously bunk ones like $1/oz, which someone probably did pay from a friend, but that doesn't reflect street value.It's hard to do this in an automated fashion, though, at least until you have a good amount of data. A good place to start would be automatically flagging figures more than 2 standard deviations from the (local) mean, and hand-curating those. Keep track of "bad" data points, and as you find patterns you can add filters.
A lot of people have been pointing on the same thing you've been pointing out lol. That's why we added the n= for the time being.
However, there are also demand and supply side things going on that we might not know about that is affecting the price.
If you cut out the user-supplied quality metric or figure out some way to try to normalize this across the samples (e.g. ask submitters if they have purchased in other states and how they would rank this purchase against that past purchase) you would probably get better useful data.
I'm not sure how practical it is to get users to try and compare product quality from different locations; the sample size would probably be very low because most people buy and consume where they live.
We could display the averages in each quantity, but it would lack enough sample size. Plus having one nice round number is easier to visualize. I think the final .oz average probably within +/-10% accuracy.