Test 2: Variable in DC for Tract

[alenastern]

I tried this for a few variables - all worked as expected on initial load.

For % White, I had issues when I clicked from map to Charts/Table and back again to map, the map showed the same reset view

There was also an issue with the scale of the chart (note the 2017/2019 entries for the tract labeled 15.5 but the bar height doesn't seem to match)

The table also doesn't seem to match the chart/map:

[Thiyaghessan]

Issue

Accordion panel tables in the Table tab display integer values like "3", "2", "1" instead of actual percentage values (e.g., "35.2%", "45.3%"). This affects Tract-level data where the underlying dataset contains duplicate rows.

Root Cause

The AFFH_tract data file contains duplicate rows for certain (geoid, year) combinations. When affh_transform() pivots the data using dcast():

data.table::dcast(dt, definition ~ year, value.var = "value")

...it encounters multiple values for the same (definition, year) cell. Without an explicit fun.aggregate parameter, dcast() defaults to length(), returning the count of rows instead of actual values.

The warning that would normally alert us to this behavior:

"Aggregate function missing, defaulting to 'length'"

...was hidden by a suppressWarnings() wrapper.

Why the Value Box Wasn't Affected

The value box uses prepare_barplot_data() which directly extracts column values without pivoting, so duplicates don't trigger aggregation.

Fix

Added deduplication before the dcast() operation in .split_affh():

affh_split <- unique(affh_split, by = c("definition", "year", "category"))

This ensures each (definition, year, category) combination has exactly one row before pivoting, so dcast() returns actual values instead of counts.

File Changed

• R/affh_transform.R - Added unique() call in .split_affh() function

Problem

This means there are duplicate rows in the tract data files. Screenshot attached.