Create pdf report of on demand ratio analyses by Damonamajor · Pull Request #21 · ccao-data/recurring-data-requests

Damonamajor · 2026-08-24T17:23:24Z

Draft can be found here:

O:\CCAODATA\tmp\ratio-analysis.pdf

Damonamajor · 2026-08-24T17:27:53Z

@wrridgeway Take a quick look at the output and see if anything stands out. After that we can do a code review.

Damonamajor · 2026-08-24T17:30:05Z

In the quick edit round, I also added a table for descriptive stats, which makes the initial text kind of obsolete. Just left in both for now.

ccao-jardine

Great work so far, @Damonamajor! Some comments here for your next iteration.

ccao-jardine · 2026-08-24T20:16:40Z

+)
+```
+
+```{r cap-1}


I like the descriptive stats table, and normally, conventions would demand it be first! But can we move the Town-Level IAAO Statistics Table up so that it is Table 1 and is the first result listed?

Reasoning:

Our stakeholders are busy and might not make it past the first page.

This table is very clear, and indicates (for the Worth output) that both the model and DR stages met IAAO stats. This is really helpful context for readers who then go on to read the rest of the report

Your intro has some helpful descriptive stats, like the total sample and population sizes.

ccao-jardine · 2026-08-24T21:00:57Z

+    df$model_value != df$desk_review_value) # nolint: indentation_linter.
+}
+
+desc_table <- tibble(


I like the idea of this table! It compels the reader to think about things they might not have thought to ask. It's very information-dense, though. Hmm. Happy to workshop this together! I'm suggesting some required changes and some optional ones below, but we can chat if that's better.

[required] First, let's relabel "No Sales" to the more conventional "Unsold". Then, in our ordering, we should always list "Sold" aud "Unsold" first, and "All Properties" last, to follow a convention of having the "total" at the very end. Finally, I think we should exclude the means in favor of medians, unless maybe there's a reason for including it I'm not thinking of?

[optional] I wonder if dividing up the data into two tables might help. What do you think about formatting it so it corresponds to answering two different questions?

Median AVs for sold and unsold properties, per stage

Model Desk Review

Sold $ $

Unsold $ $

All $ $

Desk Review changes for Sold and Unsold properties

N changes % of changes

Sold # %

Unsold # %

All # 100%

I included means since the medians often say the same (before and after DR). It shows if there was a general shift up or down in the valuations.

I also added a total AV shift to the bottom table

Ah, okay, I get your reasoning for incorporating means and medians. Visually, it's a little easier for people to read changes left to right, and here, our goal is to communicate changes (if any) from Model to DR, rather than Median to Mean.

Can we resort this such that the first column of data is Median Model AV, followed by Median DR AV, then Mean Model AV, and Mean DR AV?

ccao-jardine · 2026-08-24T21:12:31Z

+
+{{< pagebreak >}}
+
+```{r cap-2}


Let's also move our Table of Neighborhood-Level Median Ratios to be right under the map of nbhds. That way, users get to peruse stats in the following order:

Township

Deciles

NBHDs

Co-authored-by: Nicole Jardine <138712135+ccao-jardine@users.noreply.github.com>

ccao-jardine · 2026-08-25T14:05:29Z

Oh, one more request @Damonamajor: in the intro paragraph, can we list the modeling groups (Single-Family, Multi-Family, Condo?) and/or the class codes that the ratio analysis entails?

Damonamajor · 2026-08-25T20:44:46Z

Oh, one more request @Damonamajor: in the intro paragraph, can we list the modeling groups (Single-Family, Multi-Family, Condo?) and/or the class codes that the ratio analysis entails?

They come from the res model. Is that considered single and multi family?

Damonamajor · 2026-08-25T20:54:04Z

New file is in the sale shared drive. Didn't make changes on the last question since I didn't know how it should be stated.

ccao-jardine

Thanks for continuing your work on this!

ccao-jardine · 2026-08-26T22:03:16Z

+total_shift <- function(df) {
+  changed <- !is.na(df$model_value) & !is.na(df$desk_review_value) &
+    df$model_value != df$desk_review_value # nolint: indentation_linter.
+  sum(df$desk_review_value[changed] - df$model_value[changed], na.rm = TRUE)
+}


Hm, I think the total AV shift is a little hard to interpret without more context (like the total AV of the town). What about median AV (or FMV) change? Or, we can cut it.

I think we can just keep it with the median and mean MV of properties. This was just an alternative to the mean.

ccao-jardine · 2026-08-26T22:05:46Z

+```{r cap-desc-av}
+cap_desc_av <- glue(
+  "Assessed Values by Group.",
+  "\\newline Model and desk review assessed values for properties",
+  " with qualifying arm's-length sales (Sold), properties without",
+  " a recorded sale (Unsold), and all residential parcels (All)."
+)


Here and in the table, the column header says AV, but the values look like FMVs. It's important to distinguish them. Can you check these column headers? FMVs will be easier for our readers so those are preferable.

For extra clarity, it would also help to rename variable names; for example, from desk_review_value to desk_review_FMV, etc.

ccao-jardine · 2026-08-26T22:19:05Z

+print(map_diff, vp = viewport(
+  x = 2 / 3, y = 0, width = 1 / 3, height = 1,
+  just = c("left", "bottom")
+))


You know, the more I think about this third nbhd map, I think we should strike it because it might draw readers to the wrong conclusion.

Occasionally, DR will produce a higher FMV -- one that is higher than what the model produced, and much higher than the 2025 recent sale. I see this with flipped properties: they sold for a low price in 2025, got flipped, and now are significantly improved with larger square footage, new roof, etc. Often these properties are currently listed with much higher asking prices than the recent sale. Because the 2025 sale is included in the ratio analysis, this property's ratio "looks worse" because it is accurately assessed. As such, I don't want readers to draw the conclusion that it is a worse assessment.

ccao-jardine · 2026-08-26T22:22:25Z

+iaao_check <- function(metric, value) {
+  switch(metric,
+    "Median Ratio" = value >= 0.90 & value <= 1.10,
+    "COD"          = value >= 5.0 & value <= 15.0,
+    "PRD"          = value >= 0.98 & value <= 1.03,
+    "PRB"          = value >= -0.05 & value <= 0.05,
+    "MKI"          = value >= 0.90 & value <= 1.10,
+    NA
+  )
+}


I like that this is so clear! I like to think of these as global settings, and suggest the convention of moving these higher up in the quarto doc so that reviewers see these global settings before they see most of the rest of the doc. Can we move these somewhere below the input docs, and somewhere above where the Methods section starts?

ccao-jardine · 2026-08-26T22:24:28Z

+  )
+}
+
+iaao_color_cell <- function(metric, value, digits = 4) {


Same idea about the global settings: for all of these color codes, can we move these to be higher up? Can you add some comments that help the reviewer know, for example, ratio < 0.90 is blue?

ccao-jardine · 2026-08-26T22:26:02Z

+    ratio < 0.90 ~ "74ADD1",
+    ratio < 0.95 ~ "ABD9E9",
+    ratio <= 1.05 ~ "66BD63",


Let's please use the 0.90-1.10 standard for ratio, and use one shade of green to indicate when a ratio is within standards (instead of multiple shades).

ccao-jardine · 2026-08-26T22:30:59Z

+cap_fig3 <- glue(
+  "IAAO Metrics: Model vs. Desk Review.",
+  "\\newline Green shaded band shows the IAAO acceptable range for each metric."
+)
+```


Hm, thoughts on shrinking this figure of IAAO stats and moving these to be directly under Table 1 of IAAO stats?

ccao-jardine · 2026-08-26T22:33:38Z

+**`r format(n_excluded_sales, big.mark = ",")`** sales were excluded
+as outliers by Valuations analysts, leaving a **final sample of
+`r format(n_final_sales, big.mark = ",")` sales**. The township
+contains **`r format(n_total_pins, big.mark = ",")` residential


Moving our issue discussion into this PR review -- let's define what "residential" means! I'd just define it however this code does it. If it uses our residential modeling groups, let's specify which groups those are, and class codes.

Co-authored-by: Nicole Jardine <138712135+ccao-jardine@users.noreply.github.com>

…thub.com:ccao-data/recurring-data-requests into 19-create-pdf-report-of-on-demand-ratio-analyses

Damonamajor · 2026-08-28T15:47:52Z

New file is uploaded

Damonamajor · 2026-08-28T15:50:26Z

+**`r format(n_excluded_sales, big.mark = ",")`** sales were excluded
+as outliers by Valuations analysts, leaving a **final sample of
+`r format(n_final_sales, big.mark = ",")` sales**. The township
+contains **`r format(n_total_pins, big.mark = ",")` residential parcels**. This universe includes Single and Multifamily parcels (up to 6 units) and excludes condo units. Prices range from `r town_stats$price_range`.


Billy should verify this when he looks at it.

Damonamajor added 2 commits August 14, 2026 15:10

initial push

9e0d29c

post-Nicole updates

27c5429

Damonamajor linked an issue Aug 24, 2026 that may be closed by this pull request

Create pdf report of on-demand ratio analyses #19

Open

Damonamajor requested a review from ccao-jardine August 24, 2026 17:27

Damonamajor marked this pull request as ready for review August 24, 2026 17:27

Damonamajor requested a review from a team as a code owner August 24, 2026 17:27

ccao-jardine requested changes Aug 24, 2026

View reviewed changes

Damonamajor and others added 2 commits August 24, 2026 16:22

Update ratio-analysis/ratio-analysis.qmd

bc0265f

Co-authored-by: Nicole Jardine <138712135+ccao-jardine@users.noreply.github.com>

Apply suggestions from code review

e1b5ad0

Co-authored-by: Nicole Jardine <138712135+ccao-jardine@users.noreply.github.com>

Damonamajor added 3 commits August 25, 2026 17:13

ordering updates

f99714c

sepearate tables

80943a5

add total av

b114b32

Damonamajor requested a review from ccao-jardine August 25, 2026 20:53

ccao-jardine requested changes Aug 26, 2026

View reviewed changes

Damonamajor and others added 9 commits August 27, 2026 10:02

Update ratio-analysis/ratio-analysis.qmd

7440731

Co-authored-by: Nicole Jardine <138712135+ccao-jardine@users.noreply.github.com>

Nicole V2 edits

c365d96

Merge branch '19-create-pdf-report-of-on-demand-ratio-analyses' of gi…

b368006

…thub.com:ccao-data/recurring-data-requests into 19-create-pdf-report-of-on-demand-ratio-analyses

switch more to mv

aa66809

remove nsales

2192052

improve text

1cd8574

updated text

9115841

improve text

e97de6f

Add text comment

bc6eec1

Damonamajor requested a review from ccao-jardine August 28, 2026 15:47

Damonamajor commented Aug 28, 2026

View reviewed changes

Damonamajor added 5 commits August 28, 2026 16:04

small text change

65dc394

move all text

0448436

renaming of headers and cleanup

56fad5b

Fit onto one page

fbc64a4

add spacer

688f4f6

Uh oh!

Conversation

Damonamajor commented Aug 24, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Damonamajor commented Aug 24, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Damonamajor commented Aug 24, 2026

Uh oh!

ccao-jardine left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Damonamajor Aug 24, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

ccao-jardine commented Aug 25, 2026

Uh oh!

Damonamajor commented Aug 25, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Damonamajor commented Aug 25, 2026

Uh oh!

ccao-jardine left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Damonamajor commented Aug 28, 2026

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Damonamajor commented Aug 24, 2026 •

edited

Loading

Damonamajor commented Aug 24, 2026 •

edited

Loading

Damonamajor Aug 24, 2026 •

edited

Loading

Damonamajor commented Aug 25, 2026 •

edited

Loading