amended census build

analysis/GREENGridModel/_createSyntheticCensus.Rmd

@@ -51,28 +51,28 @@ Check census data distributions - we're looking for zeros.
 # do this here so we only have the relevant regions
 # > heat source ----
 
-t <- summary(censusAuWideDT[, .SD, .SDcols = names(censusAuWideDT) %like% "heat"]) # https://stackoverflow.com/questions/30189979/select-columns-of-data-table-based-on-regex
+t <- summary(au2013DT[, .SD, .SDcols = names(au2013DT) %like% "heat"]) # https://stackoverflow.com/questions/30189979/select-columns-of-data-table-based-on-regex
 
 kableExtra::kable(t, caption = "Test distribution of fuel sources") %>%
   kable_styling()
 
 # > n kids ----
 
-t <- summary(censusAuWideDT[, .SD, .SDcols = names(censusAuWideDT) %like% "nKids"])
+t <- summary(au2013DT[, .SD, .SDcols = names(au2013DT) %like% "nKids"])
 
 kableExtra::kable(t, caption = "Test distribution of nKids") %>%
   kable_styling()
 
 # > n people ----
 
-t <- summary(censusAuWideDT[, .SD, .SDcols = names(censusAuWideDT) %like% "nPeople"])
+t <- summary(au2013DT[, .SD, .SDcols = names(au2013DT) %like% "nPeople"])
 
 kableExtra::kable(t, caption = "Test distribution of nPeople") %>%
   kable_styling()
 
 # > n rooms ----
 
-t <- summary(censusAuWideDT[, .SD, .SDcols = names(censusAuWideDT) %like% "nRooms"])
+t <- summary(au2013DT[, .SD, .SDcols = names(au2013DT) %like% "nRooms"])
 
 kableExtra::kable(t, caption = "Test distribution of nRooms") %>%
   kable_styling()
@@ -82,18 +82,18 @@ kableExtra::kable(t, caption = "Test distribution of nRooms") %>%
 
 ```{r censusSetup}
 # check totals are not 0
-#censusAuWideDT <- censusAuWideDT[, nBedrooms_Total := nBedrooms_1_2 + nBedrooms_3 + nBedrooms_4m]
-censusAuWideDT <- censusAuWideDT[, nPeople_Total := nPeople_1 + nPeople_2 + nPeople_3 + nPeople_4m]
-censusAuWideDT <- censusAuWideDT[, nRooms_Total := nRooms1_4 + nRooms5_6 + nRooms7m]
-censusAuWideDT <- censusAuWideDT[, nKids_Total := nKids_0 + nKids_1m]
-censusAuWideDT <- censusAuWideDT[, heatSource_Total := heatSourceWood + heatSourceElectricity + heatSourceGas + heatSourceCoal + heatSourceOther]
+#au2013DT <- au2013DT[, nBedrooms_Total := nBedrooms_1_2 + nBedrooms_3 + nBedrooms_4m]
+au2013DT <- au2013DT[, nPeople_Total := nPeople_1 + nPeople_2 + nPeople_3 + nPeople_4m]
+au2013DT <- au2013DT[, nRooms_Total := nRooms1_4 + nRooms5_6 + nRooms7m]
+au2013DT <- au2013DT[, nKids_Total := nKids_0 + nKids_1m]
+au2013DT <- au2013DT[, heatSource_Total := heatSourceWood + heatSourceElectricity + heatSourceGas + heatSourceCoal + heatSourceOther]
 
-t <- summary(censusAuWideDT[, .SD, .SDcols = names(censusAuWideDT) %like% "_Total"])
+t <- summary(au2013DT[, .SD, .SDcols = names(au2013DT) %like% "_Total"])
 t
 message("Removing areas which have total counts for any constraint = 0")
 
-nOrig <- nrow(censusAuWideDT)
-zerosDT <- censusAuWideDT[nPeople_Total == 0 | is.na(nPeople_Total)| 
+nOrig <- nrow(au2013DT)
+zerosDT <- au2013DT[nPeople_Total == 0 | is.na(nPeople_Total)| 
                             #nBedrooms_Total == 0 | is.na(nBedrooms_Total)| 
                             nRooms_Total == 0 | is.na(nRooms_Total)| 
                             nKids_Total == 0 | is.na(nKids_Total)| 
@@ -107,7 +107,7 @@ setkey(zerosDT, AU2013_code)
 zerosDT <- auListDT[zerosDT]
 zerosDT <- zerosDT[, drop := 1]
 
-test <- zerosDT[censusAuWideDT]
+test <- zerosDT[au2013DT]
 
 # remove areas where count is 0 or NA
 censusDT <- test[is.na(drop)]
@@ -236,9 +236,12 @@ It is possible that some of the weights are 0. This means there are households w
 message("N rows before removing zero weights: ", nrow(longFormDT))
 longFormDT <- longFormDT[ipfWeight > 0]
 message("N rows after removing zero weights: ", nrow(longFormDT))
-
 ```
 
+```{r saveData}
+# save the results for future use since they won't change unless we change the constraints
+data.table::fwrite(longFormDT, paste0(sParams$ggPath, "/safe/ipf/nonZeroWeightsAu2013.csv"))
+```
 
 We now need to add the survey-based attributes back (from the GREENGrid survey). If we had a much larger sample of households and a lot more areas we would not do this here as it would create a very large file.
 
@@ -504,7 +507,7 @@ setnames(wdt, c("N.Electricity", "N.Gas", "N.Other", "N.Wood"),
 
 setkey(wdt, AU2013_code)
 
-tmpDT <- wdt[censusAuWideDT] # merge to census data
+tmpDT <- wdt[au2013DT] # merge to census data
 tmpDT <- auListDT[tmpDT] # add region labels
 
 plotDT <- tmpDT[, .(AU2013_code, REGC2013_label, AU2013_label,