Missing loss data from aal.get_summary

[aaronjamespayne]

Hello,

I am trying to compute total losses (aggregated over all floodfactor levels) for each ZCTA in 2020 and noticed that most ZCTA's only had total loss information for a subset of floodfactors. I want to know whether I can impute the missing floodfactors as having total losses equal to zero or not. For ZCTA 01001, as an example, I only managed to pull total loss information for floodfactor=4 and floodfactor=6. I know I can impute losses for floodfactor=1 as equal to zero, but I am not sure how I should impute the losses for other floodfactors, such as floodfactor=8.

In investigating this, I also noticed there were a few cases where a ZCTA had total losses equal to zero but had a positive count of properties with value loss due to flood risk. For example, ZCTA 78205, floodfactor=3, in the year 2020, has 2 properties with value loss due to flood risk under the high scenario, but the total losses under the high scenario are zero. This seems to be a contradiction since a positive count of properties with value loss should imply a positive level of total loss. Should I interpret these observations as having total loss equal to zero, or treat them as missing?

[Lyetenth]

Hi @apayne182,

We're still taking a look into the second issue, but for the first part, any floodFactors that are missing for a particular year can be assumed to be 0, ie: if floodFactor=2 is misisng, then assume both totalLoss and count for low/mid/high are all 0. I'm don't quite see the missing values for floodFactor=8 for ZCTA 01001 however, so do let me know if I'm misinterpreting the question

I'll update you on the second issue as soon as I can. Sorry for the delays

[Lyetenth]

Hi @apayne182,

Sorry for the delays. Regarding the second question with the odd cases with ZCTA AAL, this is due to a rounding decision, and the process for how how the AAL is calculated. For property AAL's on FF, we use a 3rd party AVM value to calculate the AAL which we cannot share with outside data users. That integration is dynamic and we do not hold the data in house.

In comparison, for the localities we use the internal AVM values given that they are static and do not change.
For the example ZCTA 78205, as you saw there are 2 properties under count . Their FSIDs are 4801611902 and 4801611879. If you take a look at their AAL at the property level, you can see that the AAL is $1