INGOs and instrumental factors

Code

library(tidyverse)
library(targets)
library(scales)
library(glue)
library(gt)
library(patchwork)
library(tidybayes)
library(here)

tar_config_set(store = here('_targets'),
               script = here('_targets.R'))

tar_load(c(survey_orgs, survey_countries, survey_all))
tar_load(c(models_strategies_size))

# Plotting functions
invisible(list2env(tar_read(graphic_functions), .GlobalEnv))
invisible(list2env(tar_read(table_functions), .GlobalEnv))
set_annotation_fonts()

Staffing

Code

df_staff <- survey_orgs %>%
  select(employees = Q3.4.num, volunteers = Q3.5.num) %>%
  filter(employees < 10000, volunteers < 10000) %>%
  mutate(no_emps = employees == 0,
         no_vols = volunteers == 0)

df_staff %>% 
  count(no_vols) %>% 
  mutate(prop = n / sum(n)) %>% 
  gt() %>% 
  cols_label(
    no_vols = "No volunteers",
    n = "N",
    prop = "Proportion"
  ) %>% 
  cols_align(align = "left", columns = no_vols) %>% 
  fmt_percent(columns = prop, decimals = 1) %>% 
  opts_theme()

No volunteers	N	Proportion
FALSE	537	92.6%
TRUE	43	7.4%

Code

df_staff %>% 
  group_by(no_emps, no_vols) %>% 
  summarize(across(c(employees, volunteers), 
                   list(n = ~n(), median = median, mean = mean))) %>% 
  ungroup() %>%
  mutate(prop = employees_n / sum(employees_n))
## # A tibble: 4 × 9
##   no_emps no_vols employees_n employees_median employees_mean volunteers_n volunteers_median volunteers_mean    prop
##   <lgl>   <lgl>         <int>            <dbl>          <dbl>        <int>             <dbl>           <dbl>   <dbl>
## 1 FALSE   FALSE           425                8           50.1          425                25            289. 0.733  
## 2 FALSE   TRUE             42                8           41.5           42                 0              0  0.0724 
## 3 TRUE    FALSE           112                0            0            112                20            243. 0.193  
## 4 TRUE    TRUE              1                0            0              1                 0              0  0.00172

df_staff %>% 
  pivot_longer(c(employees, volunteers)) %>% 
  mutate(name = str_to_title(name)) %>% 
  ggplot(aes(x = value, fill = name)) +
  geom_histogram(binwidth = 0.5, color = "white", linewidth = 0.25) +
  scale_x_continuous(trans = "log1p", breaks = c(0, 10^(0:5)), labels = label_comma()) +
  scale_fill_manual(values = c(clrs$Prism[3], clrs$Prism[6]), name = NULL) +
  labs(x = NULL, y = "Number of staff") +
  facet_wrap(vars(name), nrow = 2) +
  theme_ingo()

Funding

Code

df_funding <- survey_orgs %>%
  filter(!is.na(small_org)) %>%
  select(clean.id, small_org, Q3.8.num, Q3.8) %>%
  filter(!is.na(Q3.8.num))

df_funding %>%
  group_by(small_org) %>%
  summarize(across(Q3.8.num, list(mean = mean, median = median))) %>% 
  gt() %>% 
  cols_label(
    small_org = "Organization size",
    Q3.8.num_mean = "Sources (mean)",
    Q3.8.num_median = "Sources (median)"
  ) %>% 
  cols_align(align = "left", columns = small_org) %>% 
  opts_theme()

Organization size	Sources (mean)	Sources (median)
0–50 employees	3.0	3
51+ employees	3.9	4

Code

df_funding_unnested_org_size <- df_funding %>% 
  unnest(Q3.8) %>% 
  filter(key != "None") %>%
  group_by(key, value, small_org) %>%
  summarise(num = n()) %>%
  ungroup() %>%
  mutate(value = factor(value,
                        levels = levels(survey_orgs$Q3.8_individual),
                        ordered = TRUE)) %>% 
  mutate(answer = fct_collapse(value,
    "A lot" = c("A great deal", "A lot"),
    "Some" = c("A moderate amount", "A little"),
    "None" = "None at all"
  )) %>%
  filter(key != "Other") %>%
  mutate(source = fct_collapse(key,
    "NGOs and foundations" = c("Donations from other NGOs", "Foundation donations"),
    "Governments and IGOs" = c("Grants from home country", "Grants from host country", "Grants from other governments", "Grants from IGOs")
  ))

Code

df_funding_plot <- df_funding_unnested_org_size %>%
  group_by(answer, source, small_org) %>% 
  summarise(num = sum(num)) %>%
  group_by(small_org, source) %>% 
  mutate(total = sum(num)) %>% 
  ungroup() %>%
  mutate(prop = num / total) %>%
  arrange(answer, desc(small_org), desc(prop)) %>%
  mutate(source = fct_inorder(source))

df_funding_totals <- df_funding_plot %>%
  group_by(source) %>%
  summarise(total = sum(num)) %>%
  ungroup() %>%
  arrange(total) %>%
  mutate(source = fct_inorder(source))

p1 <- ggplot(df_funding_totals, aes(x = total, y = source)) +
  geom_pointrange(aes(xmin = 0, xmax = after_stat(x))) +
  scale_x_continuous(breaks = c(0, 500, 1000),
                     labels = label_comma()) +
  labs(x = "Times answer selected", y = NULL,
       title = "INGO funding sources") +
  theme_ingo() +
  theme(panel.grid.major.y = element_blank())

p2 <- ggplot(df_funding_plot, aes(x = prop, y = fct_rev(answer),
                            color = fct_rev(small_org))) +
  geom_pointrange(aes(xmin = 0, xmax = after_stat(x)),
                  position = position_dodge(width = 0.75), size = 0.5) + 
  scale_color_manual(values = c(clrs$Prism[2], clrs$Prism[5]), name = NULL,
                     guide = guide_legend(reverse = TRUE)) +
  scale_x_continuous(labels = label_percent()) +
  labs(x = "Proportion selected", y = NULL, 
       title = "How much funding INGO relies on") +
  facet_wrap(vars(source)) +
  theme_ingo() +
  theme(strip.text = element_text(size = rel(0.75)),
        panel.grid.major.y = element_blank(),
        legend.position = c(1, 0), legend.justification = c(1, 0),
        legend.direction = "vertical")

(p1 / plot_spacer() / p2) +
  plot_layout(heights = c(0.325, 0.025, 0.65))

Time in country

Code

df_time_country <- survey_all %>%
  filter(!is.na(Q4.2), Q4.2 != "Don't know") %>%
  group_by(Q4.2, target.regime.type) %>%
  summarise(num = n()) %>%
  ungroup() %>%
  mutate(Q4.2 = factor(Q4.2, levels = rev(levels(Q4.2)), ordered = TRUE))

ggplot(df_time_country, aes(x = num, y = Q4.2, color = target.regime.type)) +
  geom_pointrange(aes(xmin = 0, xmax = after_stat(x)), 
                  position = position_dodge(width = 0.5), size = 0.5) + 
  scale_x_continuous(sec.axis = sec_axis(~ . / sum(df_time_country$num),
                                         labels = label_percent())) +
  scale_color_manual(values = c(clrs$Prism[4], clrs$Prism[6]), name = NULL,
                     guide = guide_legend(reverse = TRUE)) +
  labs(x = "Number of responses", y = NULL) +
  theme_ingo() + 
  theme(panel.grid.major.y = element_blank())

Collaboration

Code

df_collaboration_raw <- survey_orgs %>%
  filter(!is.na(small_org)) %>%
  select(clean.id, small_org, Q3.6_clean) %>%
  unnest(Q3.6_clean) 

df_collaboration_raw %>%
  group_by(Q3.6_clean) %>%
  summarise(num = n()) %>%
  arrange(desc(num)) %>% 
  gt() %>% 
  cols_label(
    Q3.6_clean = "Partner",
    num = "N"
  ) %>% 
  opts_theme()

Partner	N
Other nongovernmental organizations (NGOs)	563
International organizations (IGOs)	491
Governments	420
Corporations or businesses	273
Academic organizations	45
We do not collaborate with other organizations or institutions	19
Faith-based organizations	18
Other	11
Individuals	8
Don't know	4
Media	2

Code

df_collaboration_plot <- df_collaboration_raw %>%
  filter(!(Q3.6_clean %in% c("Other", "Individuals", "Don't know", 
                             "We do not collaborate with other organizations or institutions"))) %>%
  mutate(Q3.6_clean = fct_collapse(Q3.6_clean,
    "NGOs" = 
      c("Other nongovernmental organizations (NGOs)",
        "Academic organizations",
        "Faith-based organizations"),
    "Corporations" = 
      c("Corporations or businesses", "Media"),
    "IGOs" = "International organizations (IGOs)"
  )) %>%
  group_by(small_org, Q3.6_clean) %>%
  summarise(num = n()) %>%
  group_by(small_org) %>%
  mutate(total = sum(num)) %>%
  ungroup() %>%
  mutate(prop = num / total) %>%
  arrange(small_org, prop) %>%
  mutate(Q3.6_clean = fct_inorder(Q3.6_clean))

ggplot(df_collaboration_plot, aes(x = prop, y = Q3.6_clean, 
                                  color = fct_rev(small_org))) + 
  geom_pointrange(aes(xmin = 0, xmax = after_stat(x)), 
                  position = position_dodge(width = 0.5), size = 0.5) + 
  labs(x = "Proportion selected", y = NULL) +
  scale_color_manual(values = c(clrs$Prism[2], clrs$Prism[5]), name = NULL,
                     guide = guide_legend(reverse = TRUE)) +
  scale_x_continuous(labels = label_percent()) +
  theme_ingo() + 
  theme(panel.grid.major.y = element_blank())

Flexibility and practical strategies

Size and number of changes

Code

p1 <- models_strategies_size$draws %>% 
  ggplot(aes(x = .epred, y = fct_rev(small_org), fill = small_org)) +
  stat_halfeye() +
  scale_fill_manual(values = c(clrs$Prism[2], clrs$Prism[5]),
                    guide = "none") +
  labs(x = "Average number of strategies", y = NULL) +
  theme_ingo()

p2 <- models_strategies_size$diffs %>% 
  ggplot(aes(x = .epred)) +
  stat_halfeye(fill = clrs$Prism[11]) +
  geom_vline(xintercept = 0, linewidth = 0.5, color = clrs$Prism[8]) +
  scale_x_continuous(labels = label_number(style_negative = "minus")) +
  labs(x = "Average # of strategies for 51+ employees − average for 0–50 employees",
       y = NULL) +
  theme_ingo() +
  theme(axis.text.y = element_blank(),
        panel.grid.major.y = element_blank())

(p1 / p2) +
  plot_layout(heights = c(0.75, 0.25))

Code


df_strategies_size <- survey_all %>%
  mutate(num_strategies = Q4.3_value %>% map_int(length)) %>%
  filter(!is.na(Q4.3_value), !is.na(small_org))

Code

models_strategies_size$draws %>% 
  median_qi(.epred) %>% 
  mutate(small_org = as.character(small_org),
         type = "Group medians") %>%
  rename(y = .epred, ymin = .lower, ymax = .upper) %>% 
  bind_rows(models_strategies_size$diffs_summary) %>% 
  mutate(across(c(y, ymin, ymax, p_greater_0), 
                list(nice = ~label_number(accuracy = 0.01)(.)))) %>% 
  mutate(nice_value = glue("{y_nice}<br>({ymin_nice}–{ymax_nice})")) %>% 
  mutate(small_org = str_replace(small_org, " - ", " − ")) %>% 
  select(type, small_org, nice_value, p_greater_0_nice) %>% 
  group_by(type) %>% 
  gt() %>% 
  cols_label(
    small_org = "",
    nice_value = "Posterior median",
    p_greater_0_nice = "p > 0"
  ) %>% 
  fmt_markdown(nice_value) %>% 
  sub_missing(p_greater_0_nice) %>% 
  tab_footnote(footnote = "95% credible intervals shown in parentheses") %>% 
  opts_theme()

	Posterior median	p > 0
Group medians
0–50 employees	1.32 (1.27–1.37)	—
51+ employees	1.51 (1.35–1.66)	—
Differences
51+ employees − 0–50 employees	0.19 (0.03–0.35)	0.99
95% credible intervals shown in parentheses

Collaborations, funding sources, and number of strategies

Code

df_strategies_funding <- survey_all %>%
  mutate(num_strategies = Q4.3_value %>% map_int(length)) %>%
  filter(!is.na(Q4.3_value), Q3.8.num < 7) %>%
  select(num_strategies, Q3.8.num)
  
p1 <- ggplot(df_strategies_funding, 
             aes(x = Q3.8.num, y = num_strategies, group = Q3.8.num)) +
  geom_point(position = position_jitter(seed = 1234), size = 0.5) + 
  geom_violin() +
  labs(x = "Number of funding sources", y = "Number of\noperational strategies") +
  theme_ingo()

df_strategies_collaboration <- survey_all %>%
  mutate(num_strategies = Q4.3_value %>% map_int(length),
         num_collaborations = Q3.6_num) %>%
  filter(!is.na(Q4.3_value), !is.na(Q3.6_num)) %>%
  select(num_strategies, num_collaborations)

p2 <- ggplot(df_strategies_collaboration, 
             aes(x = num_collaborations, y = num_strategies, group = num_collaborations)) +
  geom_point(position = position_jitter(seed = 1234), size = 0.5) + 
  geom_violin() +
  labs(x = "Number of collaborative partners", y = "Number of\noperational strategies") +
  theme_ingo()

p1 / p2

--- title: "INGOs and instrumental factors" format: html: code-fold: true --- ```{r setup, include=FALSE} knitr::opts_chunk$set(fig.align = "center", fig.retina = 3, fig.width = 6, fig.height = (6 * 0.618), out.width = "80%", collapse = TRUE, dev = "ragg_png") options(digits = 3, width = 120, dplyr.summarise.inform = FALSE, knitr.kable.NA = "") ``` ```{r load-libraries-data, warning=FALSE, message=FALSE} library(tidyverse) library(targets) library(scales) library(glue) library(gt) library(patchwork) library(tidybayes) library(here) tar_config_set(store = here('_targets'), script = here('_targets.R')) tar_load(c(survey_orgs, survey_countries, survey_all)) tar_load(c(models_strategies_size)) # Plotting functions invisible(list2env(tar_read(graphic_functions), .GlobalEnv)) invisible(list2env(tar_read(table_functions), .GlobalEnv)) set_annotation_fonts() ``` # Staffing ```{r tab-volunteers} df_staff <- survey_orgs %>% select(employees = Q3.4.num, volunteers = Q3.5.num) %>% filter(employees < 10000, volunteers < 10000) %>% mutate(no_emps = employees == 0, no_vols = volunteers == 0) df_staff %>% count(no_vols) %>% mutate(prop = n / sum(n)) %>% gt() %>% cols_label( no_vols = "No volunteers", n = "N", prop = "Proportion" ) %>% cols_align(align = "left", columns = no_vols) %>% fmt_percent(columns = prop, decimals = 1) %>% opts_theme() ``` ```{r tab-staff-volunteers} df_staff %>% group_by(no_emps, no_vols) %>% summarize(across(c(employees, volunteers), list(n = ~n(), median = median, mean = mean))) %>% ungroup() %>% mutate(prop = employees_n / sum(employees_n)) df_staff %>% pivot_longer(c(employees, volunteers)) %>% mutate(name = str_to_title(name)) %>% ggplot(aes(x = value, fill = name)) + geom_histogram(binwidth = 0.5, color = "white", linewidth = 0.25) + scale_x_continuous(trans = "log1p", breaks = c(0, 10^(0:5)), labels = label_comma()) + scale_fill_manual(values = c(clrs$Prism[3], clrs$Prism[6]), name = NULL) + labs(x = NULL, y = "Number of staff") + facet_wrap(vars(name), nrow = 2) + theme_ingo() ``` # Funding ```{r tab-funding} df_funding <- survey_orgs %>% filter(!is.na(small_org)) %>% select(clean.id, small_org, Q3.8.num, Q3.8) %>% filter(!is.na(Q3.8.num)) df_funding %>% group_by(small_org) %>% summarize(across(Q3.8.num, list(mean = mean, median = median))) %>% gt() %>% cols_label( small_org = "Organization size", Q3.8.num_mean = "Sources (mean)", Q3.8.num_median = "Sources (median)" ) %>% cols_align(align = "left", columns = small_org) %>% opts_theme() ``` ```{r calc-funding-org-size} df_funding_unnested_org_size <- df_funding %>% unnest(Q3.8) %>% filter(key != "None") %>% group_by(key, value, small_org) %>% summarise(num = n()) %>% ungroup() %>% mutate(value = factor(value, levels = levels(survey_orgs$Q3.8_individual), ordered = TRUE)) %>% mutate(answer = fct_collapse(value, "A lot" = c("A great deal", "A lot"), "Some" = c("A moderate amount", "A little"), "None" = "None at all" )) %>% filter(key != "Other") %>% mutate(source = fct_collapse(key, "NGOs and foundations" = c("Donations from other NGOs", "Foundation donations"), "Governments and IGOs" = c("Grants from home country", "Grants from host country", "Grants from other governments", "Grants from IGOs") )) ``` ```{r plot-funding-org-size, fig.width=7, fig.height=5} df_funding_plot <- df_funding_unnested_org_size %>% group_by(answer, source, small_org) %>% summarise(num = sum(num)) %>% group_by(small_org, source) %>% mutate(total = sum(num)) %>% ungroup() %>% mutate(prop = num / total) %>% arrange(answer, desc(small_org), desc(prop)) %>% mutate(source = fct_inorder(source)) df_funding_totals <- df_funding_plot %>% group_by(source) %>% summarise(total = sum(num)) %>% ungroup() %>% arrange(total) %>% mutate(source = fct_inorder(source)) p1 <- ggplot(df_funding_totals, aes(x = total, y = source)) + geom_pointrange(aes(xmin = 0, xmax = after_stat(x))) + scale_x_continuous(breaks = c(0, 500, 1000), labels = label_comma()) + labs(x = "Times answer selected", y = NULL, title = "INGO funding sources") + theme_ingo() + theme(panel.grid.major.y = element_blank()) p2 <- ggplot(df_funding_plot, aes(x = prop, y = fct_rev(answer), color = fct_rev(small_org))) + geom_pointrange(aes(xmin = 0, xmax = after_stat(x)), position = position_dodge(width = 0.75), size = 0.5) + scale_color_manual(values = c(clrs$Prism[2], clrs$Prism[5]), name = NULL, guide = guide_legend(reverse = TRUE)) + scale_x_continuous(labels = label_percent()) + labs(x = "Proportion selected", y = NULL, title = "How much funding INGO relies on") + facet_wrap(vars(source)) + theme_ingo() + theme(strip.text = element_text(size = rel(0.75)), panel.grid.major.y = element_blank(), legend.position = c(1, 0), legend.justification = c(1, 0), legend.direction = "vertical") (p1 / plot_spacer() / p2) + plot_layout(heights = c(0.325, 0.025, 0.65)) ``` # Time in country ```{r plot-time-in-country, fig.width=6, fig.height=3} df_time_country <- survey_all %>% filter(!is.na(Q4.2), Q4.2 != "Don't know") %>% group_by(Q4.2, target.regime.type) %>% summarise(num = n()) %>% ungroup() %>% mutate(Q4.2 = factor(Q4.2, levels = rev(levels(Q4.2)), ordered = TRUE)) ggplot(df_time_country, aes(x = num, y = Q4.2, color = target.regime.type)) + geom_pointrange(aes(xmin = 0, xmax = after_stat(x)), position = position_dodge(width = 0.5), size = 0.5) + scale_x_continuous(sec.axis = sec_axis(~ . / sum(df_time_country$num), labels = label_percent())) + scale_color_manual(values = c(clrs$Prism[4], clrs$Prism[6]), name = NULL, guide = guide_legend(reverse = TRUE)) + labs(x = "Number of responses", y = NULL) + theme_ingo() + theme(panel.grid.major.y = element_blank()) ``` # Collaboration ```{r tab-collaboration} df_collaboration_raw <- survey_orgs %>% filter(!is.na(small_org)) %>% select(clean.id, small_org, Q3.6_clean) %>% unnest(Q3.6_clean) df_collaboration_raw %>% group_by(Q3.6_clean) %>% summarise(num = n()) %>% arrange(desc(num)) %>% gt() %>% cols_label( Q3.6_clean = "Partner", num = "N" ) %>% opts_theme() ``` ```{r plot-collaboration-size, fig.width=6, fig.height=3.5} df_collaboration_plot <- df_collaboration_raw %>% filter(!(Q3.6_clean %in% c("Other", "Individuals", "Don't know", "We do not collaborate with other organizations or institutions"))) %>% mutate(Q3.6_clean = fct_collapse(Q3.6_clean, "NGOs" = c("Other nongovernmental organizations (NGOs)", "Academic organizations", "Faith-based organizations"), "Corporations" = c("Corporations or businesses", "Media"), "IGOs" = "International organizations (IGOs)" )) %>% group_by(small_org, Q3.6_clean) %>% summarise(num = n()) %>% group_by(small_org) %>% mutate(total = sum(num)) %>% ungroup() %>% mutate(prop = num / total) %>% arrange(small_org, prop) %>% mutate(Q3.6_clean = fct_inorder(Q3.6_clean)) ggplot(df_collaboration_plot, aes(x = prop, y = Q3.6_clean, color = fct_rev(small_org))) + geom_pointrange(aes(xmin = 0, xmax = after_stat(x)), position = position_dodge(width = 0.5), size = 0.5) + labs(x = "Proportion selected", y = NULL) + scale_color_manual(values = c(clrs$Prism[2], clrs$Prism[5]), name = NULL, guide = guide_legend(reverse = TRUE)) + scale_x_continuous(labels = label_percent()) + theme_ingo() + theme(panel.grid.major.y = element_blank()) ``` # Flexibility and practical strategies ## Size and number of changes ```{r plot-strategies-size} p1 <- models_strategies_size$draws %>% ggplot(aes(x = .epred, y = fct_rev(small_org), fill = small_org)) + stat_halfeye() + scale_fill_manual(values = c(clrs$Prism[2], clrs$Prism[5]), guide = "none") + labs(x = "Average number of strategies", y = NULL) + theme_ingo() p2 <- models_strategies_size$diffs %>% ggplot(aes(x = .epred)) + stat_halfeye(fill = clrs$Prism[11]) + geom_vline(xintercept = 0, linewidth = 0.5, color = clrs$Prism[8]) + scale_x_continuous(labels = label_number(style_negative = "minus")) + labs(x = "Average # of strategies for 51+ employees − average for 0–50 employees", y = NULL) + theme_ingo() + theme(axis.text.y = element_blank(), panel.grid.major.y = element_blank()) (p1 / p2) + plot_layout(heights = c(0.75, 0.25)) df_strategies_size <- survey_all %>% mutate(num_strategies = Q4.3_value %>% map_int(length)) %>% filter(!is.na(Q4.3_value), !is.na(small_org)) ``` ```{r tab-strategies-size} models_strategies_size$draws %>% median_qi(.epred) %>% mutate(small_org = as.character(small_org), type = "Group medians") %>% rename(y = .epred, ymin = .lower, ymax = .upper) %>% bind_rows(models_strategies_size$diffs_summary) %>% mutate(across(c(y, ymin, ymax, p_greater_0), list(nice = ~label_number(accuracy = 0.01)(.)))) %>% mutate(nice_value = glue("{y_nice}<br>({ymin_nice}–{ymax_nice})")) %>% mutate(small_org = str_replace(small_org, " - ", " − ")) %>% select(type, small_org, nice_value, p_greater_0_nice) %>% group_by(type) %>% gt() %>% cols_label( small_org = "", nice_value = "Posterior median", p_greater_0_nice = "p > 0" ) %>% fmt_markdown(nice_value) %>% sub_missing(p_greater_0_nice) %>% tab_footnote(footnote = "95% credible intervals shown in parentheses") %>% opts_theme() ``` ## Collaborations, funding sources, and number of strategies ```{r plot-strategies-funding-collaboration, fig.width=6, fig.height=4} df_strategies_funding <- survey_all %>% mutate(num_strategies = Q4.3_value %>% map_int(length)) %>% filter(!is.na(Q4.3_value), Q3.8.num < 7) %>% select(num_strategies, Q3.8.num) p1 <- ggplot(df_strategies_funding, aes(x = Q3.8.num, y = num_strategies, group = Q3.8.num)) + geom_point(position = position_jitter(seed = 1234), size = 0.5) + geom_violin() + labs(x = "Number of funding sources", y = "Number of\noperational strategies") + theme_ingo() df_strategies_collaboration <- survey_all %>% mutate(num_strategies = Q4.3_value %>% map_int(length), num_collaborations = Q3.6_num) %>% filter(!is.na(Q4.3_value), !is.na(Q3.6_num)) %>% select(num_strategies, num_collaborations) p2 <- ggplot(df_strategies_collaboration, aes(x = num_collaborations, y = num_strategies, group = num_collaborations)) + geom_point(position = position_jitter(seed = 1234), size = 0.5) + geom_violin() + labs(x = "Number of collaborative partners", y = "Number of\noperational strategies") + theme_ingo() p1 / p2 ```