It really is insane. It's not like we're slightly skewed towards above-average, above-average not only dominates but the above-average departures at times are just ridiculous...and even that isn't really a here and there type ordeal, it's more common than uncommon.
But what I would like to do is this:
Look at the major climo stations across the region and look at the month of July for the periods of records.
Get or compute the 30 year averages of high/low for each day of the month for each climatological period (1991-2020, 1981-2010, 1971-2000, etc.).
Create a tally of all high/low temperatures for each day of the month.
For example, July 1: how many days had highs 100+, 90-99, 80-89, 70-79, etc. Low of 80+, 70-79, 60-69, 50-59, etc.
From this data you can visualize what your lowest percentiles and highest percentiles are.
Anyways, though what I would be super curious to know is, during previous decades what was the distribution like of below-average temperatures versus above-average temperatures (and those ranges). For example, lets say average high at BDL on July 1 is 85F (which actually I think its right around there). I would like to see every single July 1 high temperature for the period of record. What I would have to wager is, if you were to look back into say the 1940's through the 1970's...what is the distribution of highs for July 1? Then compare that to the distribution of the last few decades. I wonder if there would be any clear signal.
This can be done in Python I'm sure in 15 minutes but I don't know Python and my brain isn't equip with learning it. Hell, these plots may even be able to be generated on the Iowa State page. Great site but I wish it was a little more user friendly.
Plus, its fun to crunch numbers