wikipedia_views/analysis/pageview_example.R

   1 ### COVID-19 Digital Observatory
   2 ### 2020-03-28
   3 ###
   4 ### Minimal example analysis file using pageview data
   5
   6 library(tidyverse)
   7 library(scales)
   8
   9 ### Import and cleanup one datafile from the observatory
  10
  11 DataURL <-
  12     url("https://covid19.communitydata.science/datasets/wikipedia/digobs_covid19-wikipedia-enwiki_dailyviews-20200401.tsv")
  13
  14 views <-
  15     read.table(DataURL, sep="\t", header=TRUE, stringsAsFactors=FALSE)
  16
  17 ### Alternatively, uncomment and run if working locally with full git
  18 ### tree
  19 ###
  20 ### Identify data source directory and file
  21 ## DataDir <- ("../data/")
  22 ## DataFile <- ("dailyviews2020032600.tsv")
  23
  24 ## related.searches.top <- read.table(paste(DataDir,DataFile, sep=""),
  25 ##                                   sep="\t", header=TRUE,
  26 ##                                   stringsAsFactors=FALSE)
  27
  28 ### Cleanup and do the grouping with functions from the Tidyverse
  29 ### (see https://www.tidyverse.org for more info)
  30
  31 views <- views[,c("article", "project", "timestamp", "views")]
  32 views$timestamp <- fct_explicit_na(views$timestamp)
  33
  34
  35 ### Sorts and groups at the same time
  36 views.by.proj.date <- arrange(group_by(views, project, timestamp),
  37                         desc(views))
  38
  39
  40 ### Export just the top 10 by pageviews
  41 write.table(head(views.by.proj.date, 10),
  42             file="output/top10_views_by_project_date.csv", sep=",",
  43             row.names=FALSE)
  44
  45 ### A simple visualization
  46 p <- ggplot(data=views.by.proj.date, aes(views))
  47
  48 ## Density plot with log-transformed axis
  49 p + geom_density() + scale_x_log10(labels=comma)
  50
  51
  52