Hi @laura78,
Thanks for the question. Your value of 3.25 kWh/degC does sound low. While I pointed to an example of a table with average values in the 100s range (in a previous post), values in the 10s order of magnitude are sensible, but <5 certainly sounds too low. Remember too that the expected unit for these values ought to be kWh/degC/m^2, so you’ll want to be sure to divide by the area of the building footprint. It’s essentially asking, "on average, how much energy is needed to cool each square meter of building type X, by 1 degree C?".
I like your approach and it seems rather innovative! However, I just looked a bit deeper into cooling degree days (CDD). “A degree day compares the mean… outdoor temperatures recorded for a location to a standard temperature, usually 65° Fahrenheit (F) in the United States.” I’m sure you are using degrees Celsius and not Fahrenheit, but the assumption you may have made is that the 13,550 kWh was used to cool homes to that standard temperature, 65F or ~18C. In reality, I would bet that most residents in Phoenix in July (or any month) are NOT cooling their homes all the way down to 18C. So, if you stick with your approach, using the average residential energy consumption (per square meter) for July makes sense to me, but I recommend rethinking your CDD value in the denominator. It should likely be <4,162 CDD. To obtain it for a single day, I would subtract the average residential thermostat setting in Phoenix in July (probably around 22C to 24C, or even higher) FROM Phoenix’s mean air temperature in July.
Does this make sense? If so, what order of magnitude do your residential energy consumption values end up being? I would imagine they might be approximately 20 to 30 kWh/degC/m^2. Alternatively, I am way off and/or you thought about all that already.
Please let us know. (I wanted to send you a reply sooner than later 
, but someone from our Cities team may also chime-in with more intimate knowledge of this model.)
Be well,
Jesse