Condo descriptives (mean and std) update

[TimCookCountyDS]

Adds mean and std to the condo summary stats. (See the gen_agg_stats function in pipeline/03-evaluate.R)- Closes issue #147 .
Matches the condo descriptives to those of the Res-val model. (ccao-data/model-res-avm#458).

New outputs can be found in the model performance tables:
(Adding mean and sd to both lists produces 16 new columns in total):

From sum_fns_list:

┌─────────────────────────────────────────────┐
│ New column │
├─────────────────────────────────────────────┤
│ sale_fmv_mean / sale_fmv_sd │
├─────────────────────────────────────────────┤
│ prior_far_fmv_mean / prior_far_fmv_sd │
├─────────────────────────────────────────────┤
│ prior_near_fmv_mean / prior_near_fmv_sd │
├─────────────────────────────────────────────┤
│ estimate_fmv_mean / estimate_fmv_sd │
└─────────────────────────────────────────────┘

From sum_sqft_fns_list:

┌───────────────────────────────────────────────────────────────┐
│ New column │
├───────────────────────────────────────────────────────────────┤
│ sale_fmv_per_sqft_mean / sale_fmv_per_sqft_sd │
├───────────────────────────────────────────────────────────────┤
│ prior_far_fmv_per_sqft_mean / prior_far_fmv_per_sqft_sd │
├───────────────────────────────────────────────────────────────┤
│ prior_near_fmv_per_sqft_mean / prior_near_fmv_per_sqft_sd │
├───────────────────────────────────────────────────────────────┤
│ estimate_fmv_per_sqft_mean / estimate_fmv_per_sqft_sd │
└───────────────────────────────────────

[wrridgeway]

@TimCookCountyDS 3 things will i'm looking at this

• rather than link to the issue in the PR description, you can link it to this PR using the "development" drop down on the right side, that way the PR closes the issue when it's merged
• can we remove your name from the PR title (Tim Condo is a great nom de plume)
• can we get pre-commit passing

[jeancochrane]

rather than link to the issue in the PR description, you can link it to this PR using the "development" drop down on the right side, that way the PR closes the issue when it's merged

Just adding that I like to do both of these things (noting the issue in the comment and linking the PR to the issue using the "development" dropdown)! Though I agree with Billy that linking the issue is most important in cases where there's a 1:1 relationship between a PR and an issue.

[wrridgeway]

Thanks for adding this for condos!

[wrridgeway]

Is this documenting something that's actually happening? If it is, could we clean this formatting up?

[TimCookCountyDS]

O weird, I thought I removed that before I pushed

[TimCookCountyDS]

It was just a note to self, will cut it out

[TimCookCountyDS]

Removed

[TimCookCountyDS]

@TimCookCountyDS 3 things will i'm looking at this

• rather than link to the issue in the PR description, you can link it to this PR using the "development" drop down on the right side, that way the PR closes the issue when it's merged
• can we remove your name from the PR title (Tim Condo is a great nom de plume)
• can we get pre-commit passing

Will do- for pre-commit: when I run pre-commit on the server (before I push), I use pre-commit run. Should I be using a different command?

[wrridgeway]

For me after I've installed the pre-commit git hooks (pre-commit install) for the repo I don't even need to run pre-commit. git commit should trigger pre-commit, and you shouldn't be able to commit without passing.

Is that right @jeancochrane ?

[TimCookCountyDS]

For me after I've installed the pre-commit git hooks (pre-commit install) for the repo I don't even need to run pre-commit. git commit should trigger pre-commit, and you shouldn't be able to commit without passing.

Is that right @jeancochrane ?

got - it- done :)

[TimCookCountyDS]

@TimCookCountyDS 3 things will i'm looking at this

• rather than link to the issue in the PR description, you can link it to this PR using the "development" drop down on the right side, that way the PR closes the issue when it's merged
  done
• can we remove your name from the PR title (Tim Condo is a great nom de plume)
  done
• can we get pre-commit passing
  done (as far as I can tell)

--- squash and merge, @wrridgeway ?

[jeancochrane]

For me after I've installed the pre-commit git hooks (pre-commit install) for the repo I don't even need to run pre-commit. git commit should trigger pre-commit, and you shouldn't be able to commit without passing.

Correct! We mention this in our pre-commit docs, though perhaps it could be called out more explicitly, since I've noticed it's a step that may people miss: https://github.com/ccao-data/wiki/blob/master/How-To/Use-pre-commit-to-lint-and-format-your-code.md#installing-pre-commit

[TimCookCountyDS]

For me after I've installed the pre-commit git hooks (pre-commit install) for the repo I don't even need to run pre-commit. git commit should trigger pre-commit, and you shouldn't be able to commit without passing.

Correct! We mention this in our pre-commit docs, though perhaps it could be called out more explicitly, since I've noticed it's a step that may people miss: https://github.com/ccao-data/wiki/blob/master/How-To/Use-pre-commit-to-lint-and-format-your-code.md#installing-pre-commit

For me after I've installed the pre-commit git hooks (pre-commit install) for the repo I don't even need to run pre-commit. git commit should trigger pre-commit, and you shouldn't be able to commit without passing.

Correct! We mention this in our pre-commit docs, though perhaps it could be called out more explicitly, since I've noticed it's a step that may people miss: https://github.com/ccao-data/wiki/blob/master/How-To/Use-pre-commit-to-lint-and-format-your-code.md#installing-pre-commit

Thanks! I think my mental model was slightly off when I read those-

[TimCookCountyDS]

@Damonamajor would you mind running this with build-and-run model and giving the outputs a quick once over? (I did + checked that the additional columns were added + populated for the new model run- see comment above for new columns and gave the model report a quick once over)- but would you mind a quick spot check just in case I missed anything?

(not that I expect any problematic changes- but good to have a 2nd set of eyes on it before I merge)

[Damonamajor]

https://github.com/ccao-data/model-condo-avm/actions/runs/24792538947

I presume we expect to see the changes in model.stats (or something like this) on aws and in the performance report?

[TimCookCountyDS]

https://github.com/ccao-data/model-condo-avm/actions/runs/24792538947

I presume we expect to see the changes in model.stats (or something like this) on aws and in the performance report?

I didn't update the performance reports yet- so the main place I've been keeping track of the changes is updates to the model.performance table -
See comment for a list of the new columns #153 (comment)

If mean and std should also be propagating to stats (and aren't) lmk, and we can think through it - (Mean and std currently follow the exact logic of the other summary stats functions, so wherever those show up, mean and std should as well).

[wrridgeway]

Seems fine to merge after Damon has confirmed the columns appear as expected.

[Damonamajor]

Looks good to me. Just for reference these are the columns which are in the table for

SELECT * FROM "model"."performance"
where run_id = '2026-04-21-sharp-manasi'

but not in

2026-02-11-recursing-rob

sale_fmv_mean | sale_fmv_sd | sale_fmv_per_sqft_mean | sale_fmv_per_sqft_sd | prior_far_fmv_mean | prior_far_fmv_sd | prior_far_fmv_per_sqft_mean | prior_far_fmv_per_sqft_sd | prior_near_fmv_mean | prior_near_fmv_sd | prior_near_fmv_per_sqft_mean | prior_near_fmv_per_sqft_sd | estimate_fmv_mean | estimate_fmv_sd | estimate_fmv_per_sqft_mean | estimate_fmv_per_sqft_sd

[Damonamajor]

https://github.com/ccao-data/model-condo-avm/actions/runs/24792538947
I presume we expect to see the changes in model.stats (or something like this) on aws and in the performance report?

I didn't update the performance reports yet- so the main place I've been keeping track of the changes is updates to the model.performance table - See comment for a list of the new columns #153 (comment)

If mean and std should also be propagating to stats (and aren't) lmk, and we can think through it - (Mean and std currently follow the exact logic of the other summary stats functions, so wherever those show up, mean and std should as well).

Model.stats doesn't exist. It's what I was guessing model.performance was called.

[Damonamajor]

Looks good to me