cleaner approach to managing the data with Git

.gitignore

@@ -35,6 +35,4 @@ vignettes/*.pdf
 
 
 
-webrear.zip
-data/all-data.csv
 *.xlsx

build.R

@@ -23,6 +23,6 @@ source("prep/download-mrtes.R")
 # load in the data (not really necessary)
 load("data/mrtes.rda")
 
-# buidl the report.  Note that the build will happen in the user's home directory (~).
+# build the actual report.  Note that the build will happen in the user's home directory (~).
 build_doc(subdir = "report-1")
 

data/readme.txt

@@ -0,0 +1 @@
+holds both original data (eg .xlsx) and persistent versions of R data (.rda).  But these are ignored by Git, this is just for your local clone.

prep/download-mrtes.R

@@ -4,12 +4,14 @@
 url <- "http://www.mbie.govt.nz/info-services/sectors-industries/tourism/tourism-research-data/monthly-regional-tourism-estimates/document-image-library/international-product-by-rto.xlsx"
 
 # Only download the file if we haven't already done it as I don't want to cause unnecessary network traffic
-if(!file.exists("mrtes.xlsx")){
-  download.file(url, destfile = "mrtes.xlsx", mode = "wb")
+if(!file.exists("data/mrtes.xlsx")){
+  download.file(url, destfile = "data/mrtes.xlsx", mode = "wb")
 }
 
-mrtes_orig <- read.xlsx("mrtes.xlsx", sheet = "Data base", detectDates = TRUE)
+mrtes_orig <- read.xlsx("data/mrtes.xlsx", sheet = "Data base", detectDates = TRUE)
 
+# make some alternative names for the Regional Tourism Organisations, more grpahic friendly,
+# including one that is just the 5 big ones plus "Other":
 mrtes <- mrtes_orig %>%
   mutate(rto_short = gsub(" RTO", "", RTO),
          rto_lumped = fct_reorder(rto_short, -Spend),

report-1/report.Rmd

@@ -44,7 +44,13 @@ We see a very strong seasonal pattern, and also that people spend a lot on accom
 
 ```{r}
 p <- mrtes %>%
+  # make products friendly with line wraps for the facet titles:
   mutate(Product = str_wrap(Product, 20)) %>%
+  # add up all the "others""
+  group_by(Date, Product, rto_lumped) %>%
+  summarise(Spend = sum(Spend)) %>%
+  ungroup() %>%
+  mutate(rto_lumped = fct_reorder(rto_lumped, -Spend)) %>%
   ggplot(aes(x = Date, y = Spend, colour = rto_lumped)) +
   facet_wrap(~Product, scales = "free_y") +
   theme(legend.position = c(0.9, 0.1)) +