renaming example analysis directories

keywords/example_analysis/output/top5_queries_per_term_per_date.csv

@@ -0,0 +1,11 @@
+"term","date","query.1","query.2","query.3","query.4","query.5"
+"coronavirus",2020-03-27,coronavirus update,corona,coronavirus symptoms,news coronavirus,coronavirus cases
+"covid-19",2020-03-27,covid-19 coronavirus,coronavirus,covid,covid-19 cases,covid 19
+"covid-19 pandemic",2020-03-27,coronavirus,covid-19 coronavirus pandemic,coronavirus pandemic,who,is covid-19 a pandemic
+"covid19",2020-03-27,covid,covid 19,coronavirus covid19,coronavirus,covid19 cases
+"sars-cov-2",2020-03-27,coronavirus,coronavirus sars-cov-2,covid-19,covid-19 sars-cov-2,sars
+"coronavirus",2020-03-28,coronavirus update,corona,coronavirus symptoms,news coronavirus,coronavirus cases
+"covid-19",2020-03-28,coronavirus,coronavirus covid-19,covid,covid-19 cases,covid 19
+"covid-19 pandemic",2020-03-28,coronavirus pandemic,coronavirus,covid-19 coronavirus pandemic,is covid-19 a pandemic,who pandemic
+"covid19",2020-03-28,covid,covid 19,coronavirus covid19,coronavirus,covid19 cases
+"sars-cov-2",2020-03-28,coronavirus sars-cov-2,coronavirus,sars-cov-2 covid-19,covid-19,sars

keywords/example_analysis/related_searches_example.R

@@ -0,0 +1,36 @@
+### COVID-19 Digital Observatory
+### 2020-03-28
+### 
+### Minimal example analysis file using trending search data
+
+### Import and cleanup data
+
+DataURL <-
+    url("https://github.com/CommunityDataScienceCollective/COVID-19_Digital_Observatory/blob/master/transliterations/data/output/related_searches_top.csv")
+
+related.searches.top <- read.table(DataURL,
+                           sep=",", header=TRUE,
+                           stringsAsFactors=FALSE) 
+
+### Alternatively, uncomment and run if working locally with full git tree
+### Identify data source directory and file
+## DataDir <- ("../data/output/")
+## DataFile <- ("related_searches_top.csv")
+
+## related.searches.top <- read.table(paste(DataDir,DataFile, sep=""),
+##                                   sep=",", header=TRUE,
+##                                   stringsAsFactors=FALSE)
+
+### Aggregate top 5 search queries by term/day
+top5.per.term.date <- aggregate(query ~ term + date,
+                                data=related.searches.top,
+                                head, 5)
+
+## Might cleanup a bit for further analysis or visualization...
+top5.per.term.date$date <- as.Date(top5.per.term.date$date)
+
+### Export
+write.table(top5.per.term.date,
+            file="output/top5_queries_per_term_per_date.csv", sep=",",
+            row.names=FALSE)
+

keywords/example_analysis/related_searches_tidyverse_example.R

@@ -0,0 +1,31 @@
+### COVID-19 Digital Observatory
+### 2020-03-28
+### 
+### Minimal example analysis file using trending search data
+
+library(tidyverse)
+
+### Import and cleanup data
+
+
+related.searches.top = read_csv("https://github.com/CommunityDataScienceCollective/COVID-19_Digital_Observatory/raw/master/keywords/output/intermediate/related_searches_top.csv")
+
+
+## Plot how often the top 10 queries appear in the top 10 suggested list each day
+
+plot <- related.searches.top %>% 
+  group_by(term, date) %>% # Group by term and date
+  arrange(-value) %>% # Sort by value (this should already be done anyway)
+  top_n(10) %>% # Get the top 10 queries for each term-day pair
+  group_by(query) %>% # Group by again, this time for each query
+  summarize(appearances = n()) %>% # Count how often this query appears in the top 10 (which is how many Google displays)
+  arrange(-appearances) %>% # Sort by appearances
+  top_n(10) %>% # And get the top 10 queries
+  ggplot(aes(x=reorder(query, appearances), y=appearances)) + # Plot the number of appearances, ordered by appearances
+  geom_bar(stat = 'identity') + # Tell R that we want to use the values of `appearances` as the counts
+  coord_flip() + # Flip the plot
+  xlab("Query") + 
+  ylab("Number of appearances in top 10 suggested queries") +
+  theme_minimal() # And make it minimal
+
+ggsave('./output/top_queries_plot.png', plot)

keywords/example_analysis/translations_example.R

@@ -0,0 +1,17 @@
+## example reading latest file straight from the server
+df <- read.csv("https://covid19.communitydata.science/datasets/keywords/csv/latest.csv")
+
+## make the data more R-friendly
+df$is.alt <- df$is_alt == "True"
+df$is_alt <- NULL
+
+## find all translations for coronavirus
+coronavirus.itemids <- df[ (tolower(df$label) == "coronavirus") &
+                         (df$langcode == 'en')
+                       ,"itemid"]
+
+## there are actually 5 item ids. The one referring to the family of virus is Q57751738
+coronavirus.translations <- df[df$itemid == "http://www.wikidata.org/entity/Q57751738",]
+
+## let's only look at non-aliases
+print(coronavirus.translations[c(coronavirus.translations$is.alt == FALSE), c("label","langcode")])

keywords/example_analysis/translations_example.py

@@ -0,0 +1,13 @@
+import pandas as pd
+
+# read the latest dataset
+df  = pd.read_csv("https://covid19.communitydata.science/datasets/keywords/csv/latest.csv")
+
+# find translations of "coronavirus"
+coronavirus_itemids = df.loc[df.label.str.lower() == "coronavirus"]
+
+# there are actually 5 item ids. The one referring to the family of virus is Q57751738
+coronavirus_translations = df.loc[df.itemid == "http://www.wikidata.org/entity/Q57751738"]
+
+# let's only look at unique, non-aliases
+print(coronavirus_translations.loc[df.is_alt == False,['label','langcode']])