From: Aaron Shaw <aaronshaw@northwestern.edu>
Date: Wed, 1 Apr 2020 15:41:02 +0000 (-0500)
Subject: Merge pull request #11 from jdfoote/master
X-Git-Url: https://code.communitydata.science/covid19.git/commitdiff_plain/4f8a698c62f878e248dc1c4d4bb8e048a79fb661?hp=8bb3db8b46e22311a5b7a1b0f88c8cc84c649699

Merge pull request #11 from jdfoote/master

Adding a tidyverse example (with very verbose comments)
---

diff --git a/keywords/analysis/output/top_queries_plot.png b/keywords/analysis/output/top_queries_plot.png
new file mode 100644
index 0000000..276610e
Binary files /dev/null and b/keywords/analysis/output/top_queries_plot.png differ
diff --git a/keywords/analysis/related_searches_tidyverse_example.R b/keywords/analysis/related_searches_tidyverse_example.R
new file mode 100644
index 0000000..f22ab0d
--- /dev/null
+++ b/keywords/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)
\ No newline at end of file