Active transportation

R/active-transportation.R

@@ -146,11 +146,11 @@ active_transport2_fun <-
 
   }
 
-#' @title Daily active transportation (2015-2018)
+#' @title Daily active transportation (2015-2020)
 #' 
 #' @description This function creates a derived variable for daily time spent 
 #' traveling in active ways. This includes walking and biking. This function is 
-#' used for CCHS 2015-2018.
+#' used for CCHS 2015-2020.
 #' 
 #' @param PAYDVTTR number of minutes of active transportation in a week 
 #' for 12-17 years old.
@@ -162,7 +162,7 @@ active_transport2_fun <-
 #' 
 #' @examples
 #' # Using active_transport3_fun() to determine daily time spent 
-#' # traveling in active ways values across CCHS 2015-2018.
+#' # traveling in active ways values across CCHS 2015-2020.
 #' 
 #' # active_transport3_fun() is specified in variable_details.csv along with the CCHS
 #' # variables and cycles included.

ceps/cep-016-active-transport/PR-176-review-summary.md

@@ -0,0 +1,62 @@
+# PR #176 review summary: active transportation
+
+## Scope
+
+- **Variables**: PAC_4A, PAC_4A_cont, PAC_4B, PAC_4B_cont, PAC_7, PAC_7A, PAC_7B, PAC_7B_cont, PAC_8, PAC_8A, PAC_8B, PAC_8B_cont, PAYDVTTR, PAADVTRV, active_transport (15 variables)
+- **Database types**: PUMF (`_p`) and Master (`_m`)
+- **Cycles**: 2001 through 2019-2020
+- **Verification**: Three-way triangulation (MCP + Gem + Claude Code)
+
+## Checks performed
+
+### L0-L2: documentation review
+
+- Source variable names verified via MCP cchs-metadata for all 3 eras
+- Pre-2007 cycle letters confirmed: PACA_4A (2001), PACC_4A (2003), PACE_4A (2005)
+- PAYDVTTR/PAADVTRV confirmed in 2015-2016 through 2019-2020 (PUMF + Master)
+- Both also exist in 2021 Master (expansion opportunity, not flagged as issue)
+- Response categories verified against MCP value codes
+
+### L3-L5: worksheet checks
+
+- Era boundary defaults: correct — each era uses appropriate `[VAR]` defaults
+- databaseStart: consistent between variables.csv and variable_details.csv
+- PUMF/Master naming: identical source names (no split needed)
+- Pre-2007 cycle letters: correct db::VAR mappings in variable_details.csv
+- No known error patterns found
+
+### L6: implementation validation
+
+- `rec_with_table()` ran successfully for all PUMF cycles (cchs2001_p through cchs2017_2018_p)
+- Direct recode variables: valid% consistent within eras
+- active_transport DV: tested with explicit feeders per era
+- Era 1 (2001-2005): ~98% valid (hours-based, broad coverage)
+- Era 2 (2007-2014): ~60-71% valid (expected — gate question routes non-workers to NA)
+- Era 3 (2015-2016, 2017-2018): ~98% valid (StatCan derived, age-routed)
+- No step changes at era boundaries beyond expected design differences
+
+### Gem cross-check
+
+- All 4 tiers confirmed clean by NotebookLM Gem against ~239 StatCan PDFs
+- One Gem false positive: claimed PAADVTRV range was [0,6300] but worksheet correctly has [0,10080]. Gem confused PAYDVTTR (youth, [0,6300]) with PAADVTRV (adult, [0,10080]). Confirmed via MCP.
+
+## Issues found and fixed
+
+1. **PAC_4B_cont labelLong** (pre-existing): said "walking" instead of "biking" in `variables.csv`. Fixed in this commit.
+
+### Already clean (verified)
+
+- **`_s` databases**: The PR author already removed deprecated `_s` suffixes from active_transport rows. No `_s` references remain in any in-scope variable.
+
+### Informational notes (not blocking)
+
+1. **2021 Master expansion**: PAYDVTTR and PAADVTRV exist in cchs2021 Master per MCP — candidate for future addition
+
+## Artifacts
+
+- `integration-test-active-transport.R` — L6 test for direct recode variables
+- `integration-test-active-transport-dv.R` — L6 test for active_transport DV
+- `active-transport-integration-test.csv` — L6 results
+- `gn-active-transport-prompt.md` — Gem cross-check prompt
+- `gn-active-transport-variables.csv` — filtered variables.csv rows
+- `gn-active-transport-variable_details.csv` — filtered variable_details.csv rows

ceps/cep-016-active-transport/active-transport-integration-test.csv

@@ -0,0 +1,69 @@
+"era","cycle","variable","n","valid","valid_pct"
+"ERA1","cchs2001_p","PAC_4A",200,200,100
+"ERA1","cchs2001_p","PAC_4A_cont",200,197,98.5
+"ERA1","cchs2001_p","PAC_4B",200,200,100
+"ERA1","cchs2001_p","PAC_4B_cont",200,197,98.5
+"ERA1","cchs2003_p","PAC_4A",200,200,100
+"ERA1","cchs2003_p","PAC_4A_cont",200,193,96.5
+"ERA1","cchs2003_p","PAC_4B",200,200,100
+"ERA1","cchs2003_p","PAC_4B_cont",200,195,97.5
+"ERA1","cchs2005_p","PAC_4A",200,200,100
+"ERA1","cchs2005_p","PAC_4A_cont",200,195,97.5
+"ERA1","cchs2005_p","PAC_4B",200,200,100
+"ERA1","cchs2005_p","PAC_4B_cont",200,197,98.5
+"ERA2","cchs2007_2008_p","PAC_7",200,200,100
+"ERA2","cchs2007_2008_p","PAC_7A",200,37,18.5
+"ERA2","cchs2007_2008_p","PAC_7B",200,200,100
+"ERA2","cchs2007_2008_p","PAC_7B_cont",200,37,18.5
+"ERA2","cchs2007_2008_p","PAC_8",200,200,100
+"ERA2","cchs2007_2008_p","PAC_8A",200,7,3.5
+"ERA2","cchs2007_2008_p","PAC_8B",200,200,100
+"ERA2","cchs2007_2008_p","PAC_8B_cont",200,7,3.5
+"ERA2","cchs2009_2010_p","PAC_7",200,200,100
+"ERA2","cchs2009_2010_p","PAC_7A",200,24,12
+"ERA2","cchs2009_2010_p","PAC_7B",200,200,100
+"ERA2","cchs2009_2010_p","PAC_7B_cont",200,24,12
+"ERA2","cchs2009_2010_p","PAC_8",200,200,100
+"ERA2","cchs2009_2010_p","PAC_8A",200,3,1.5
+"ERA2","cchs2009_2010_p","PAC_8B",200,200,100
+"ERA2","cchs2009_2010_p","PAC_8B_cont",200,3,1.5
+"ERA2","cchs2010_p","PAC_7",200,200,100
+"ERA2","cchs2010_p","PAC_7A",200,26,13
+"ERA2","cchs2010_p","PAC_7B",200,200,100
+"ERA2","cchs2010_p","PAC_7B_cont",200,26,13
+"ERA2","cchs2010_p","PAC_8",200,200,100
+"ERA2","cchs2010_p","PAC_8A",200,9,4.5
+"ERA2","cchs2010_p","PAC_8B",200,200,100
+"ERA2","cchs2010_p","PAC_8B_cont",200,9,4.5
+"ERA2","cchs2011_2012_p","PAC_7",200,200,100
+"ERA2","cchs2011_2012_p","PAC_7A",200,28,14
+"ERA2","cchs2011_2012_p","PAC_7B",200,200,100
+"ERA2","cchs2011_2012_p","PAC_7B_cont",200,28,14
+"ERA2","cchs2011_2012_p","PAC_8",200,200,100
+"ERA2","cchs2011_2012_p","PAC_8A",200,10,5
+"ERA2","cchs2011_2012_p","PAC_8B",200,200,100
+"ERA2","cchs2011_2012_p","PAC_8B_cont",200,10,5
+"ERA2","cchs2012_p","PAC_7",200,200,100
+"ERA2","cchs2012_p","PAC_7A",200,26,13
+"ERA2","cchs2012_p","PAC_7B",200,200,100
+"ERA2","cchs2012_p","PAC_7B_cont",200,26,13
+"ERA2","cchs2012_p","PAC_8",200,200,100
+"ERA2","cchs2012_p","PAC_8A",200,5,2.5
+"ERA2","cchs2012_p","PAC_8B",200,200,100
+"ERA2","cchs2012_p","PAC_8B_cont",200,5,2.5
+"ERA2","cchs2013_2014_p","PAC_7",200,200,100
+"ERA2","cchs2013_2014_p","PAC_7A",200,26,13
+"ERA2","cchs2013_2014_p","PAC_7B",200,200,100
+"ERA2","cchs2013_2014_p","PAC_7B_cont",200,26,13
+"ERA2","cchs2013_2014_p","PAC_8",200,200,100
+"ERA2","cchs2013_2014_p","PAC_8A",200,4,2
+"ERA2","cchs2013_2014_p","PAC_8B",200,200,100
+"ERA2","cchs2013_2014_p","PAC_8B_cont",200,4,2
+"ERA2","cchs2014_p","PAC_7",200,200,100
+"ERA2","cchs2014_p","PAC_7A",200,30,15
+"ERA2","cchs2014_p","PAC_7B",200,200,100
+"ERA2","cchs2014_p","PAC_7B_cont",200,30,15
+"ERA2","cchs2014_p","PAC_8",200,200,100
+"ERA2","cchs2014_p","PAC_8A",200,2,1
+"ERA2","cchs2014_p","PAC_8B",200,200,100
+"ERA2","cchs2014_p","PAC_8B_cont",200,2,1

ceps/cep-016-active-transport/gn-active-transport-prompt.md

@@ -0,0 +1,75 @@
+# Active transportation variable cross-check
+
+## Context
+
+I'm reviewing cchsflow PR #176 which adds active transportation variables to the harmonization worksheets. cchsflow harmonizes CCHS variables across cycles using `variable_details.csv` (recoding rules) and `variables.csv` (variable registry).
+
+Two CSV files are attached:
+- `gn-active-transport-variables.csv` — 15 variables from variables.csv
+- `gn-active-transport-variable_details.csv` — 97 rows from variable_details.csv
+
+## Variable inventory
+
+The variables fall into 3 eras:
+
+**Era 1 (2001-2005): Hours per week**
+- PAC_4A / PAC_4A_cont — Hours walking to work/school (6-category → midpoint)
+- PAC_4B / PAC_4B_cont — Hours biking to work/school (6-category → midpoint)
+- Source names: PACA_4A (2001), PACC_4A (2003), PACE_4A (2005) — same for 4B
+
+**Era 2 (2007-2014): Gate + frequency + duration**
+- PAC_7 — Walked to work/school? (Yes/No/No work)
+- PAC_7A — Number of times walked (continuous, [1,270])
+- PAC_7B / PAC_7B_cont — Time per walk (4-category → midpoint in minutes)
+- PAC_8 — Biked to work/school? (Yes/No/No work)
+- PAC_8A — Number of times biked (continuous, [1,270])
+- PAC_8B / PAC_8B_cont — Time per bike (4-category → midpoint in minutes)
+
+**Era 3 (2015-2020): StatCan derived minutes per week**
+- PAYDVTTR — Total minutes active transportation, 7 days, youth 12-17 (continuous)
+- PAADVTRV — Active transportation minutes per week, adult 18+ (continuous)
+
+**Derived variable:**
+- active_transport — Daily active transportation (minutes), computed by 3 era-specific functions
+
+## Review scope — please verify
+
+### Tier 1: Database coverage
+For each variable, check whether the CCHS cycles listed in `databaseStart` are correct:
+
+1. **PAC_4A/4B**: Listed for cchs2001_p/m, cchs2003_p/m, cchs2005_p/m only. Confirm these variables (PACA_4A, PACC_4A, PACE_4A and equivalents for 4B) exist in those cycles and do NOT exist in 2007+.
+
+2. **PAC_7/7A/7B and PAC_8/8A/8B**: Listed for cchs2007_2008 through cchs2013_2014 (both _p and _m), plus single-year databases cchs2010_p, cchs2012_p, cchs2014_p, cchs2009_m, cchs2010_m, cchs2012_m. Confirm:
+   - PAC_7/8 exist in 2007-2014 cycles
+   - PAC_7/8 do NOT exist in 2015+ (replaced by PAYDVTTR/PAADVTRV)
+   - The single-year databases (cchs2009_m, cchs2010_m, cchs2012_m, cchs2010_p, cchs2012_p, cchs2014_p) are valid
+
+3. **PAYDVTTR/PAADVTRV**: Listed for cchs2015_2016 through cchs2019_2020 (both _p and _m). Confirm:
+   - These derived variables exist in 2015-2020 cycles
+   - Whether they also exist in 2021+ cycles (expansion opportunity)
+   - Whether 2019-2020 PUMF has these variables
+
+4. **active_transport**: DerivedVar spanning all 3 eras plus Master databases. The PR newly adds: cchs2001_m through cchs2019_2020_m and cchs2019_2020_p. Confirm feeder variables exist in those databases.
+
+### Tier 2: Source variable name mappings
+1. Pre-2007 cycle letters: Is PACA_4A correct for 2001? PACC_4A for 2003? PACE_4A for 2005? Same pattern for 4B?
+2. For 2007-2014: Is the source variable name simply PAC_7, PAC_7A, PAC_7B, PAC_8, PAC_8A, PAC_8B (no rename)?
+3. For 2015+: Were PAYDVTTR and PAADVTRV introduced in 2015? Any renames in later cycles?
+
+### Tier 3: Response categories
+1. **PACA_4A/4B** (2001-2005): Confirm 6 categories: 1=None, 2=<1hr, 3=1-5hrs, 4=6-10hrs, 5=11-20hrs, 6=>20hrs. Missing: 96=NA, 97-99=DK/Ref/NS.
+2. **PAC_7/8** (2007-2014): Confirm 3 categories: 1=Yes, 2=No, 3=No work/school. Missing: 6=NA, 7-9=DK/Ref/NS.
+3. **PAC_7B/8B** (2007-2014): Confirm 4 categories: 1=1-15min, 2=16-30min, 3=31-60min, 4=>1hr. Missing: 6=NA, 7-9=DK/Ref/NS.
+4. **PAC_7A/8A** (2007-2014): Continuous [1,270]. Missing: 996=NA, 997-999=DK/Ref/NS.
+5. **PAYDVTTR**: Continuous. What is the valid range? Missing codes?
+6. **PAADVTRV**: Continuous. What is the valid range? Missing codes?
+
+### Tier 4: Derived variable logic
+1. The `active_transport` variable uses three era-specific functions. For Era 3 (2015-2020), the function adds PAYDVTTR and PAADVTRV then divides by 7. Given that PAYDVTTR is youth-only and PAADVTRV is adult-only (age-routed), is it correct that a respondent would have only one non-missing value?
+
+## Output format
+
+For each tier, please report:
+- **Confirmed**: Items verified against your sources
+- **Issues**: Discrepancies found (with source citation)
+- **Unable to verify**: Items not in your source documents