### COVID-19 Digital Observatory
### 2020-03-28
### 
### Minimal example analysis file using pageview data

library(tidyverse)
library(scales)

### Import and cleanup one datafile from the observatory

DataURL <-
    url("https://covid19.communitydata.science/datasets/wikipedia/digobs_covid19-wikipedia-enwiki_dailyviews-20200101.tsv")

views <-
    read.table(DataURL, sep="\t", header=TRUE, stringsAsFactors=FALSE) 

### Cleanup and do the grouping with functions from the Tidyverse
### (see https://www.tidyverse.org for more info)

views <- views[,c("article", "project", "timestamp", "views")]
views$timestamp <- fct_explicit_na(as.character(views$timestamp))


### Sorts and groups at the same time
views.by.proj.date <- arrange(group_by(views, project, timestamp),
                        desc(views))

### Export just the top 10 by pageviews
write.table(head(views.by.proj.date, 10),
            file="output/top10_views_by_project_date.csv", sep=",",
            row.names=FALSE)

### A simple visualization
p <- ggplot(data=views.by.proj.date, aes(views))

## Density plot with log-transformed axis
p + geom_density() + scale_x_log10(labels=comma)