Figure updates

- Corrected small translation issues in date scales
- Added new panels to Supp. Fig. 3 showing views and downloads in month of preprint posting

figures.Rmd

@@ -457,7 +457,6 @@ p2E <- preprints %>%
   scale_x_date(date_breaks = "1 month",
                date_minor_breaks = "1 month",
                date_labels = "%b") +
-  theme(axis.text.x = element_text(angle = 45, hjust = 1)) +
   ggsave("outputs/figures/partials/Figure_2E.png", width = 10, height = 4)
 
 # Panel F: Reference counts
@@ -482,7 +481,6 @@ p2F <- preprints %>%
   scale_x_date(date_breaks = "1 month",
                date_minor_breaks = "1 month",
                date_labels = "%b") +
-  theme(axis.text.x = element_text(angle = 45, hjust = 1)) +
   ggsave("outputs/figures/partials/Figure_2F.png", width = 10, height = 4)
 
 p <- p2A + p2B + p2C + p2D + p2E + p2F +
@@ -654,17 +652,18 @@ p3B <- preprints %>%
          posted_date <= analysis_end,
          institution_match_country_code %in% top_n_countries) %>%
   count(institution_match_country_code, covid_preprint) %>%
-  group_by(institution_match_country_code) %>%
-  mutate(total_n = sum(n)) %>%
+  group_by(covid_preprint) %>%
+  mutate(total_n = sum(n),
+         proportion = (n/sum(n))*100) %>%
   ungroup() %>%
   mutate(continent = countrycode::countrycode(institution_match_country_code, 
                                               origin = 'iso2c', 
                                               destination = 'continent')) %>%
   assign_covid_preprint() %>%
-  ggplot(aes(x = reorder(institution_match_country_code, -n), y = n, fill = covid_preprint)) +
+  ggplot(aes(x = reorder(institution_match_country_code, -n), y = proportion, fill = covid_preprint)) +
   geom_col(color = "grey50", position = position_dodge(preserve = "single"), 
            width = 0.75, size = 0.25) +
-  labs(x = "Country of corresponding author", y = "Preprints deposited", fill = "") +
+  labs(x = "Country of corresponding author", y = "% of preprints", fill = "") +
   scale_fill_manual(values = qualitative_hcl(2, palette = "Set2")) +
   theme(legend.position = "top") +
   ggsave("outputs/figures/partials/Figure_3B.png", width = 5, height = 4)
@@ -874,8 +873,7 @@ p4A <- preprints %>%
   scale_x_date(date_breaks = "1 month",
                date_minor_breaks = "1 month",
                date_labels = "%b") +
-  theme(axis.text.x = element_text(angle = 45, hjust = 1),
-        legend.position = "top") +
+  theme(legend.position = "top") +
   ggsave("outputs/figures/partials/Figure_4A.png", width = 5, height = 4)
 
 # Panel B: Journal locations of published papers (top 10)
@@ -1005,7 +1003,6 @@ p5A <- d %>%
   scale_x_date(date_breaks = "1 month",
                date_minor_breaks = "1 month",
                date_labels = "%b") +
-  theme(axis.text.x = element_text(angle = 45, hjust = 1)) +
   ggsave("outputs/figures/partials/Figure_5A.png", width = 10, height = 4)
 
 # PDF downloads
@@ -1025,7 +1022,6 @@ p5B <- d %>%
   scale_x_date(date_breaks = "1 month",
                date_minor_breaks = "1 month",
                date_labels = "%b") +
-  theme(axis.text.x = element_text(angle = 45, hjust = 1)) +
   ggsave("outputs/figures/partials/Figure_5B.png", width = 10, height = 4)
 
 # Patchwork
@@ -1039,7 +1035,59 @@ p5A + p5B + plot_layout(ncol = 1) +
 
 ```{r}
 
-# Panel A: Abstract views (bioRxiv + medRxiv, non-COVID)
+# Panel A: Abstract views in first month after posting
+d <- preprint_usage %>%
+  mutate(
+    collection_month = str_sub(collection_date, 1, 3),
+    collection_year = str_c("20", str_sub(collection_date, 
+                                          nchar(collection_date) - 1,
+                                          nchar(collection_date))),
+    collection_date = my(str_c(collection_month, collection_year))) %>%
+  inner_join(preprints, by = "doi") %>%
+  assign_covid_preprint() %>%
+  mutate(posted_month = floor_date(posted_date, unit = "month")) %>%
+  filter(posted_month >= analysis_start,
+         posted_month <= analysis_end,
+         posted_month == collection_date)
+
+pS3A <- d %>%
+  ggplot(aes(x = posted_month, y = abstract_views, color = covid_preprint,
+             group = interaction(posted_month, covid_preprint))) +
+  geom_jitter(aes(fill = factor(covid_preprint)),
+              shape = 21, size = 0.25, alpha = 0.1, 
+              position = position_jitterdodge(jitter.width = 15, dodge.width = 21)) +
+  geom_boxplot(width = 20, outlier.shape = NA, fill = NA,
+               position = position_dodge(width = 21)) +
+  labs(x = "Posted Month", y = "Abstract views in posting month", 
+       fill = "", color = "") +
+  scale_y_log10(labels = scales::comma, limits = c(1, 1e6), expand = c(0, 0)) +
+  scale_color_manual(values = qualitative_hcl(2, palette = "Set2")) +
+  scale_fill_manual(values = qualitative_hcl(2, palette = "Set2")) +
+  scale_x_date(date_breaks = "1 month",
+               date_minor_breaks = "1 month",
+               date_labels = "%b") +
+  ggsave("outputs/figures/partials/Figure_S3A.png", width = 10, height = 4)
+
+# Panel B: PDF downloads
+pS3B <- d %>%
+  ggplot(aes(x = posted_month, y = pdf_downloads, color = covid_preprint,
+             group = interaction(posted_month, covid_preprint))) +
+  geom_jitter(aes(fill = factor(covid_preprint)),
+              shape = 21, size = 0.25, alpha = 0.1, 
+              position = position_jitterdodge(jitter.width = 15, dodge.width = 21)) +
+  geom_boxplot(width = 20, outlier.shape = NA, fill = NA,
+               position = position_dodge(width = 21)) +
+  labs(x = "Posted Month", y = "PDF downloads in posting month", 
+       fill = "", color = "") +
+  scale_y_log10(labels = scales::comma, limits = c(1, 1e6), expand = c(0, 0)) +
+  scale_color_manual(values = qualitative_hcl(2, palette = "Set2")) +
+  scale_fill_manual(values = qualitative_hcl(2, palette = "Set2")) +
+  scale_x_date(date_breaks = "1 month",
+               date_minor_breaks = "1 month",
+               date_labels = "%b") +
+  ggsave("outputs/figures/partials/Figure_S3B.png", width = 10, height = 4)
+
+# Panel C: Abstract views (bioRxiv + medRxiv, non-COVID)
 d <- preprint_usage %>%
   inner_join(preprints, by = "doi") %>%
   filter(covid_preprint != T) %>%
@@ -1052,7 +1100,7 @@ d <- preprint_usage %>%
   filter(posted_date >= "2019-01-01",
          posted_date <= analysis_end)
 
-pS3A <- d %>%
+pS3C <- d %>%
   ggplot(aes(x = posted_month, y = abstract_views, group = posted_month)) +
   geom_boxplot(outlier.shape = NA) +
   geom_jitter(shape = 21, size = 0.2, alpha = 0.1) +
@@ -1066,11 +1114,10 @@ pS3A <- d %>%
                date_labels = "%b-%y",
                expand = c(0.01,0.01)) +
   theme(axis.text.x = element_text(angle = 45, hjust = 1)) +
-  ggsave("outputs/figures/partials/Figure_S3A.png", width = 10, height = 4)
-
-# Panel B: PDF downloads (bioRxiv + medRxiv, non-COVID)
+  ggsave("outputs/figures/partials/Figure_S3C.png", width = 10, height = 4)
 
-pS3B <- d %>%
+# Panel D: PDF downloads (bioRxiv + medRxiv, non-COVID)
+pS3D <- d %>%
   ggplot(aes(x = posted_month, y = pdf_downloads, group = posted_month)) +
   geom_boxplot(outlier.shape = NA) +
   geom_jitter(shape = 21, size = 0.2, alpha = 0.1) +
@@ -1084,9 +1131,9 @@ pS3B <- d %>%
                date_labels = "%b-%y",
                expand = c(0.01,0.01)) +
   theme(axis.text.x = element_text(angle = 45, hjust = 1)) +
-  ggsave("outputs/figures/partials/Figure_S3B.png", width = 10, height = 4)
+  ggsave("outputs/figures/partials/Figure_S3D.png", width = 10, height = 4)
 
-# Panel C: PDF downloads for additional preprint servers
+# Panel E: PDF downloads for additional preprint servers
 # Bind into one data frame
 all_server_dloads <- bind_rows(preprint_usage %>%
                                  inner_join(preprints, by = c("doi", "source")) %>%
@@ -1099,7 +1146,7 @@ all_server_dloads <- bind_rows(preprint_usage %>%
   filter(posted_date >= analysis_start & posted_date <= analysis_end)
 
 # Generate plot
-pS3C <- all_server_dloads %>%
+pS3E <- all_server_dloads %>%
   assign_covid_preprint() %>%
   ggplot(aes(x = source, y = pdf_downloads, color = covid_preprint)) +
   geom_boxplot(outlier.shape = NA) +
@@ -1111,14 +1158,13 @@ pS3C <- all_server_dloads %>%
   scale_y_log10(labels = scales::comma, limits = c(1, 1e7), expand = c(0, 0)) +
   scale_color_manual(values = qualitative_hcl(2, palette = "Set2")) +
   scale_fill_manual(values = qualitative_hcl(2, palette = "Set2")) +
-  theme(legend.position = "top") + 
-  ggsave("outputs/figures/partials/Figure_S3C.png", width = 10, height = 4)
+  ggsave("outputs/figures/partials/Figure_S3E.png", width = 10, height = 4)
 
 # Patchwork
-pS3A + pS3B + pS3C +
-  plot_layout(nrow = 3) +
+pS3A + pS3B + pS3C + pS3D + pS3E +
+  plot_layout(nrow = 5) +
   plot_annotation(tag_levels = "A") +
-  ggsave("outputs/figures/Figure_S3.png", width = 10, height = 8)
+  ggsave("outputs/figures/Figure_S3.png", width = 10, height = 14)
 
 ```
 
@@ -1136,7 +1182,7 @@ d <- preprints %>%
                                   week_start = 1)) %>%
   assign_covid_preprint()
 
-plot_main <- function(d, title_y) {
+plot_main <- function(d, title_y, limit_y) {
   
   d %>%
     group_by(covid_preprint, posted_week) %>%
@@ -1146,7 +1192,7 @@ plot_main <- function(d, title_y) {
     geom_line() +
     geom_point(shape = 21) +
     labs(x = "Date posted", y = title_y, fill = "", color = "") +
-    scale_y_continuous(labels = scales::percent, limits = c(0,1)) +
+    scale_y_continuous(labels = scales::percent, limits = c(0,limit_y)) +
     scale_color_manual(values = qualitative_hcl(2, palette = "Set2")) +
     scale_fill_manual(values = qualitative_hcl(2, palette = "Set2"))  +
     guides(fill = FALSE) +
@@ -1180,7 +1226,8 @@ plot_inset <- function(d, title_y) {
 # Citations
 p6A_main <- d %>%
   mutate(is_cited = as.numeric(citations > 0)) %>%
-  plot_main(title_y = str_wrap("% of preprints with \u2265 1 citation", 30))
+  plot_main(title_y = str_wrap("% of preprints with \u2265 1 citation", 30),
+            limit_y = 1)
 p6A_inset <- d %>%
   mutate(metric = citations) %>%
   plot_inset(title_y = "Citations")
@@ -1194,7 +1241,8 @@ p6A <- p6A_main +
 # Twitter
 p6B_main <- d %>%
   mutate(is_cited = as.numeric(twitter > 1)) %>%
-  plot_main(title_y = str_wrap("% of preprints with \u2265 2 tweets*", 30))
+  plot_main(title_y = str_wrap("% of preprints with \u2265 2 tweets*", 30),
+            limit_y = 1)
 p6B_inset <- d %>%
   mutate(metric = twitter) %>%
   plot_inset(title_y = "Tweets")
@@ -1208,7 +1256,8 @@ p6B <- p6B_main +
 # News
 p6C_main <- d %>%
   mutate(is_cited = as.numeric(news > 0)) %>%
-  plot_main(title_y = str_wrap("% of preprints mentioned in \u2265 1 news article", 30))
+  plot_main(title_y = str_wrap("% of preprints mentioned in \u2265 1 news article", 30),
+            limit_y = 1)
 p6C_inset <- d %>%
   mutate(metric = news) %>%
   plot_inset(title_y = "News")
@@ -1222,7 +1271,8 @@ p6C <- p6C_main +
 # Blogs
 p6D_main <- d %>%
   mutate(is_cited = as.numeric(blogs > 0)) %>%
-  plot_main(title_y = str_wrap("% of preprints mentioned in \u2265 1 blog post", 30))
+  plot_main(title_y = str_wrap("% of preprints mentioned in \u2265 1 blog post", 30),
+            limit_y = 1)
 p6D_inset <- d %>%
   mutate(metric = blogs) %>%
   plot_inset(title_y = "Blogs")
@@ -1236,21 +1286,23 @@ p6D <- p6D_main +
 # Wikipedia
 p6E_main <- d %>%
   mutate(is_cited = as.numeric(wikipedia > 0)) %>%
-  plot_main(title_y = str_wrap("% of preprints with \u2265 1 Wikipedia citation", 30))
+  plot_main(title_y = str_wrap("% of preprints with \u2265 1 Wikipedia citation", 30),
+            limit_y = 0.1)
 p6E_inset <- d %>%
   mutate(metric = wikipedia) %>%
   plot_inset("Wikipedia")
 p6E <- p6E_main + 
   annotation_custom(ggplotGrob(p6E_inset), 
                     xmin = as.Date("2020-06-01"), 
                     xmax = as.Date("2020-11-07"), 
-                    ymin = 0.4, ymax = 1.04) +
+                    ymin = 0.04, ymax = 0.104) +
   ggsave("outputs/figures/partials/Figure_6E.png", width = 5, height = 4)
 
 # Comments
 p6F_main <- d %>%
   mutate(is_cited = as.numeric(comments_count > 0)) %>%
-  plot_main(title_y = str_wrap("% of preprints with \u2265 1 comment", 30))
+  plot_main(title_y = str_wrap("% of preprints with \u2265 1 comment", 30),
+            limit_y = 1)
 p6F_inset <- d %>%
   mutate(metric = comments_count) %>%
   plot_inset(title_y = "Comments")