In MarkEdmondson1234/googleCloudRunner: R Scripts in the Google Cloud via Cloud Run, Cloud Build and Cloud Scheduler

knitr::opts_chunk$set(echo = TRUE)

googleCloudRunner - combining R with other languages example

This is an Rmd file that is downloaded within the Cloud Build polygot demo as described on the googleCloudRunner use cases.

Previous Cloud Build steps

The previous steps have download an encrypted authentication file, decrypted it and used gago to download Google Analytics data.

The gago docker image was called for these arguments:

gagocli reports --view=81416156 --dims='ga:date,ga:medium' --mets='ga:sessions' --start=2014-01-01 --end=2019-11-30 -v --max=-1 -o=google_analytics.csv

R processing

This data is now available to be sourced by this Rmd file in google_analytics.csv

ga <- read.csv("google_analytics.csv", stringsAsFactors = FALSE)
knitr::kable(head(ga))

Any R processes can be done with an Rmd file, so long as the calling R image has the correct packages installed - this one is gcr.io/gcer-public/packagetools:latest which has tidyverse and other package tools installed.

For this demo, we decompose the data to tease out seasonality and trends:

library(tidyverse)
pivot <- ga %>% 
  spread(ga.medium, ga.sessions) %>%
  arrange(ga.date)

pivot_ts <- ts(pivot[, c("referral","(none)","organic")], frequency = 30)
plot(pivot_ts, axes = FALSE)
pivot_ts[is.na(pivot_ts)] <- 0
plot(decompose(pivot_ts[, "organic"]), axes = FALSE)

Post Cloud Build steps

This Rmd file is then rendered into HTML that is hosted on Cloud Run. It is intended as a demo of setting up a scheduled reporting service, whilst using many applications and services besides R

MarkEdmondson1234/googleCloudRunner documentation built on Feb. 5, 2023, 5:45 p.m.