[MediaWiki-commits] [Gerrit] wikimedia...rainbow[develop]: Add sister search traffic

2017-06-28 Thread Chelsyx (Code Review)
Chelsyx has submitted this change and it was merged. ( 
https://gerrit.wikimedia.org/r/361902 )

Change subject: Add sister search traffic
..


Add sister search traffic

- Adds a "Sister Search" section with a "Traffic" subsection

Bug: T164854
Change-Id: Ic89b51f3b89b25b50387389ef84ba9496423be4b
---
M server.R
A tab_documentation/sister_search_traffic.md
M ui.R
M utils.R
4 files changed, 91 insertions(+), 13 deletions(-)

Approvals:
  Chelsyx: Verified; Looks good to me, approved



diff --git a/server.R b/server.R
index 0de1586..92e94c1 100644
--- a/server.R
+++ b/server.R
@@ -18,20 +18,22 @@
 read_desktop()
 progress$set(message = "Downloading apps data", value = 0.1)
 read_apps()
-progress$set(message = "Downloading mobile web data", value = 0.3)
+progress$set(message = "Downloading mobile web data", value = 0.2)
 read_web()
-progress$set(message = "Downloading API usage data", value = 0.4)
+progress$set(message = "Downloading API usage data", value = 0.3)
 read_api()
-progress$set(message = "Downloading zero results data", value = 0.5)
+progress$set(message = "Downloading zero results data", value = 0.4)
 read_failures()
-progress$set(message = "Downloading engagement data", value = 0.6)
+progress$set(message = "Downloading engagement data", value = 0.5)
 read_augmented_clickthrough()
-progress$set(message = "Downloading language-project engagement data", 
value = 0.7)
+progress$set(message = "Downloading language-project engagement data", 
value = 0.6)
 read_augmented_clickthrough_langproj()
-progress$set(message = "Downloading survival data", value = 0.8)
+progress$set(message = "Downloading survival data", value = 0.7)
 read_lethal_dose()
-progress$set(message = "Downloading PaulScore data", value = 0.9)
+progress$set(message = "Downloading PaulScore data", value = 0.8)
 read_paul_score()
+progress$set(message = "Downloading sister search data", value = 0.9)
+read_sister_search()
 progress$set(message = "Finished downloading datasets", value = 1)
 existing_date <<- Sys.Date()
 progress$close()
@@ -359,6 +361,40 @@
   dyEvent(as.Date("2017-01-01"), "R (reportupdater)", labelLoc = "bottom")
   })
 
+  # Sister Search
+  output$sister_search_traffic_plot <- renderDygraph({
+switch(
+  input$sister_search_traffic_split,
+  "project" = {
+sister_search_traffic %>%
+  dplyr::rename(split = project)
+  },
+  "destination" = {
+sister_search_traffic %>%
+  dplyr::mutate(split = dplyr::if_else(is_serp, "Search results page", 
"Article"))
+  },
+  "language" = {
+sister_search_traffic %>%
+  dplyr::filter(project != "wikimedia commons", !is.na(language)) %>%
+  dplyr::mutate(split = language)
+  },
+  "access_method" = {
+sister_search_traffic %>%
+  dplyr::mutate(split = access_method)
+  }
+) %>%
+  dplyr::group_by(date, split) %>%
+  dplyr::summarize(pageviews = sum(pageviews)) %>%
+  tidyr::spread(split, pageviews, fill = 0) %>%
+  polloi::smoother(smooth_level = 
polloi::smooth_switch(input$smoothing_global, 
input$smoothing_sister_search_traffic_plot)) %>%
+  polloi::make_dygraph(xlab = "Date", ylab = "Pageviews", title = "Traffic 
to sister projects from Wikipedia SERPs") %>%
+  dyAxis("x", ticker = "Dygraph.dateTicker", axisLabelFormatter = 
polloi::custom_axis_formatter,
+ axisLabelWidth = 100, pixelsPerLabel = 80) %>%
+  dyLegend(labelsDiv = "sister_search_traffic_plot_legend") %>%
+  dyRangeSelector(fillColor = "", strokeColor = "") %>%
+  dyEvent(as.Date("2017-06-15"), "A (deployed)", labelLoc = "bottom")
+  })
+
   # Survival
   output$lethal_dose_plot <- renderDygraph({
 user_page_visit_dataset %>%
diff --git a/tab_documentation/sister_search_traffic.md 
b/tab_documentation/sister_search_traffic.md
new file mode 100644
index 000..6258a6b
--- /dev/null
+++ b/tab_documentation/sister_search_traffic.md
@@ -0,0 +1,28 @@
+Sister search traffic
+===
+Sister (cross-wiki) search is a feature that adds results from other projects 
to a sidebar on the search engine results page (SERP). For example: if there 
are additional results found, users are shown images from Wikimedia Commons, 
definitions from Wiktionary, and results from works on Wikisource. See 
[T146667](https://phabricator.wikimedia.org/T146667) for more details.
+
+Notes
+-
+Some communities (e.g. Italian Wikipedia) developed their own cross-wiki 
search results sidebars, which is why we see some sister traffic before the 
deployment of the sister search feature across all Wikipedias.
+
+__\*__ Users can click on a cross-wiki result or view all the results at the 
sister project
+
+__†__ This excludes the language-less Wikimedia Commons
+
+Outages and inaccuracies
+--
+* '__A

[MediaWiki-commits] [Gerrit] wikimedia...rainbow[develop]: Add sister search traffic

2017-06-28 Thread Bearloga (Code Review)
Bearloga has uploaded a new change for review. ( 
https://gerrit.wikimedia.org/r/361902 )

Change subject: Add sister search traffic
..

Add sister search traffic

- Adds a "Sister Search" section with a "Traffic" subsection

Bug: T164854
Change-Id: Ic89b51f3b89b25b50387389ef84ba9496423be4b
---
M server.R
A tab_documentation/sister_search_traffic.md
M ui.R
M utils.R
4 files changed, 91 insertions(+), 13 deletions(-)


  git pull ssh://gerrit.wikimedia.org:29418/wikimedia/discovery/rainbow 
refs/changes/02/361902/1

diff --git a/server.R b/server.R
index 0de1586..92e94c1 100644
--- a/server.R
+++ b/server.R
@@ -18,20 +18,22 @@
 read_desktop()
 progress$set(message = "Downloading apps data", value = 0.1)
 read_apps()
-progress$set(message = "Downloading mobile web data", value = 0.3)
+progress$set(message = "Downloading mobile web data", value = 0.2)
 read_web()
-progress$set(message = "Downloading API usage data", value = 0.4)
+progress$set(message = "Downloading API usage data", value = 0.3)
 read_api()
-progress$set(message = "Downloading zero results data", value = 0.5)
+progress$set(message = "Downloading zero results data", value = 0.4)
 read_failures()
-progress$set(message = "Downloading engagement data", value = 0.6)
+progress$set(message = "Downloading engagement data", value = 0.5)
 read_augmented_clickthrough()
-progress$set(message = "Downloading language-project engagement data", 
value = 0.7)
+progress$set(message = "Downloading language-project engagement data", 
value = 0.6)
 read_augmented_clickthrough_langproj()
-progress$set(message = "Downloading survival data", value = 0.8)
+progress$set(message = "Downloading survival data", value = 0.7)
 read_lethal_dose()
-progress$set(message = "Downloading PaulScore data", value = 0.9)
+progress$set(message = "Downloading PaulScore data", value = 0.8)
 read_paul_score()
+progress$set(message = "Downloading sister search data", value = 0.9)
+read_sister_search()
 progress$set(message = "Finished downloading datasets", value = 1)
 existing_date <<- Sys.Date()
 progress$close()
@@ -359,6 +361,40 @@
   dyEvent(as.Date("2017-01-01"), "R (reportupdater)", labelLoc = "bottom")
   })
 
+  # Sister Search
+  output$sister_search_traffic_plot <- renderDygraph({
+switch(
+  input$sister_search_traffic_split,
+  "project" = {
+sister_search_traffic %>%
+  dplyr::rename(split = project)
+  },
+  "destination" = {
+sister_search_traffic %>%
+  dplyr::mutate(split = dplyr::if_else(is_serp, "Search results page", 
"Article"))
+  },
+  "language" = {
+sister_search_traffic %>%
+  dplyr::filter(project != "wikimedia commons", !is.na(language)) %>%
+  dplyr::mutate(split = language)
+  },
+  "access_method" = {
+sister_search_traffic %>%
+  dplyr::mutate(split = access_method)
+  }
+) %>%
+  dplyr::group_by(date, split) %>%
+  dplyr::summarize(pageviews = sum(pageviews)) %>%
+  tidyr::spread(split, pageviews, fill = 0) %>%
+  polloi::smoother(smooth_level = 
polloi::smooth_switch(input$smoothing_global, 
input$smoothing_sister_search_traffic_plot)) %>%
+  polloi::make_dygraph(xlab = "Date", ylab = "Pageviews", title = "Traffic 
to sister projects from Wikipedia SERPs") %>%
+  dyAxis("x", ticker = "Dygraph.dateTicker", axisLabelFormatter = 
polloi::custom_axis_formatter,
+ axisLabelWidth = 100, pixelsPerLabel = 80) %>%
+  dyLegend(labelsDiv = "sister_search_traffic_plot_legend") %>%
+  dyRangeSelector(fillColor = "", strokeColor = "") %>%
+  dyEvent(as.Date("2017-06-15"), "A (deployed)", labelLoc = "bottom")
+  })
+
   # Survival
   output$lethal_dose_plot <- renderDygraph({
 user_page_visit_dataset %>%
diff --git a/tab_documentation/sister_search_traffic.md 
b/tab_documentation/sister_search_traffic.md
new file mode 100644
index 000..6258a6b
--- /dev/null
+++ b/tab_documentation/sister_search_traffic.md
@@ -0,0 +1,28 @@
+Sister search traffic
+===
+Sister (cross-wiki) search is a feature that adds results from other projects 
to a sidebar on the search engine results page (SERP). For example: if there 
are additional results found, users are shown images from Wikimedia Commons, 
definitions from Wiktionary, and results from works on Wikisource. See 
[T146667](https://phabricator.wikimedia.org/T146667) for more details.
+
+Notes
+-
+Some communities (e.g. Italian Wikipedia) developed their own cross-wiki 
search results sidebars, which is why we see some sister traffic before the 
deployment of the sister search feature across all Wikipedias.
+
+__\*__ Users can click on a cross-wiki result or view all the results at the 
sister project
+
+__†__ This excludes the language-less Wikimedia Commons
+
+Outages a