rtraining 1.2.5

DESCRIPTION

@@ -1,6 +1,6 @@
 Package: rtraining
 Title: R Training
-Version: 1.2.4
+Version: 1.2.5
 Authors@R:
     person("John", "Benninghoff", , "jbenninghoff@mac.com", role = c("aut", "cre"),
            comment = c(ORCID = "0000-0002-6230-4742"))

NEWS.md

@@ -1,3 +1,7 @@
+# rtraining 1.2.5
+
+* Updated Constraints vs Performance: Added "Transparent Donut" section
+
 # rtraining 1.2.4
 
 * Updated Constraints vs Performance: Added "Growth of Controls" section

_freeze/analysis/constraints/execute-results/html.json

@@ -1,8 +1,8 @@
 {
-  "hash": "358f18699c0ed68bbe1560c938b92ea9",
+  "hash": "3152510323d5e2b19e5f0cb5589617f0",
   "result": {
     "engine": "knitr",
-    "markdown": "---\ntitle: \"Constraints vs Performance\"\nauthor: \"John Benninghoff\"\ndate: '2024-04-30'\ndate-modified: '2024-04-30'\ncategories: notes\norder: 107\noutput:\n  html_notebook:\n    theme:\n      version: 5\n      preset: bootstrap\n    css: assets/extra.css\n    pandoc_args: --shift-heading-level-by=1\n    toc: yes\n    toc_float:\n      collapsed: no\n      smooth_scroll: no\n---\n\n\nVisualizations exploring the use of constraints vs performance improvements in risk management.\n\n# Questions/TODO\n\n- [ ] Questions/TODO list here\n\n\n::: {.cell}\n\n```{.r .cell-code}\nlibrary(ggplot2)\nlibrary(jbplot)\nlibrary(fs)\nlibrary(tibble)\nlibrary(dplyr)\n```\n:::\n\n\n# Normal Performance\n\nReplicate a version of Figure 9 from the\n[Safety-II White Paper](https://www.england.nhs.uk/signuptosafety/wp-content/uploads/sites/16/2015/10/safety-1-safety-2-whte-papr.pdf),\nwith help from <https://ggplot2tutor.com/tutorials/sampling_distributions>:\n\n\n::: {.cell}\n\n```{.r .cell-code}\nxmin <- -5\nxmax <- 5\n\nsave_png <- function(filename) {\n  ggsave(filename = path(\"rendered\", filename), width = 16 * 0.6, height = 9 * 0.6, bg = \"white\")\n}\n\nbackground <- ggplot(data.frame(x = c(xmin, xmax)), aes(x)) +\n  scale_x_continuous(breaks = -3:3, minor_breaks = NULL) +\n  labs(x = NULL, y = NULL) +\n  theme_quo(minor.y = FALSE)\n\nbaseline <- stat_function(fun = dnorm, geom = \"line\")\nbad <- stat_function(fun = dnorm, geom = \"area\", fill = \"red\", xlim = c(xmin, -2))\n\nbackground + bad + baseline\n```\n\n::: {.cell-output-display}\n![](constraints_files/figure-html/background-1.png){width=816}\n:::\n\n```{.r .cell-code}\nsave_png(\"01-baseline-bad.png\")\n```\n:::\n\n\nThe plot above shows \"bad\" outcomes in red. Let's add in \"good\" outcomes (>1) in green:\n\n\n::: {.cell}\n\n```{.r .cell-code}\ngood <- stat_function(fun = dnorm, geom = \"area\", fill = \"green\", xlim = c(1, xmax))\n\nbackground + bad + good + baseline\n```\n\n::: {.cell-output-display}\n![](constraints_files/figure-html/good-1.png){width=816}\n:::\n\n```{.r .cell-code}\nsave_png(\"02-baseline-bad-good.png\")\n```\n:::\n\n\n# Constrained Performance\n\nOne way of reducing \"bad\" outcomes is by constraining performance - reducing the standard\ndeviation.\n\n\n::: {.cell}\n\n```{.r .cell-code}\nconstrained <- stat_function(fun = dnorm, args = list(sd = 0.7), geom = \"line\", color = \"blue\")\ntaller <- scale_y_continuous(limits = c(0, 0.6))\n\nbackground +\n  stat_function(\n    fun = dnorm, args = list(sd = 0.7), geom = \"area\", fill = \"red\", xlim = c(xmin, -2)\n  ) +\n  stat_function(\n    fun = dnorm, args = list(sd = 0.7), geom = \"area\", fill = \"green\", xlim = c(1, xmax)\n  ) +\n  constrained +\n  taller\n```\n\n::: {.cell-output-display}\n![](constraints_files/figure-html/constrained-1.png){width=816}\n:::\n\n```{.r .cell-code}\nsave_png(\"03-constrained.png\")\n```\n:::\n\n\nPlotting both on the same grid shows the reduction in both \"bad\" and \"good\" outcomes:\n\n\n::: {.cell}\n\n```{.r .cell-code}\nbackground +\n  bad +\n  good +\n  baseline +\n  constrained +\n  taller\n```\n\n::: {.cell-output-display}\n![](constraints_files/figure-html/overlay-1.png){width=816}\n:::\n\n```{.r .cell-code}\nsave_png(\"04-baseline-constrained.png\")\n```\n:::\n\n\n# Improved Performance\n\nAnother way of reducing bad outcomes is by improving performance - shifting the mean.\n\n\n::: {.cell}\n\n```{.r .cell-code}\nperformance <- stat_function(fun = dnorm, args = list(mean = 1), geom = \"line\", color = \"blue\")\nimproved <- stat_function(\n  fun = dnorm, args = list(mean = 1), geom = \"area\", fill = \"green\", xlim = c(1, xmax)\n)\n\nbackground +\n  stat_function(\n    fun = dnorm, args = list(mean = 1), geom = \"area\", fill = \"red\", xlim = c(xmin, -2)\n  ) +\n  improved +\n  performance\n```\n\n::: {.cell-output-display}\n![](constraints_files/figure-html/performance-1.png){width=816}\n:::\n\n```{.r .cell-code}\nsave_png(\"05-improved.png\")\n```\n:::\n\n\nPlotting both together shows a reduction in \"bad\" and an increase in \"good\" outcomes:\n\n\n::: {.cell}\n\n```{.r .cell-code}\nbackground +\n  bad +\n  improved +\n  baseline +\n  performance\n```\n\n::: {.cell-output-display}\n![](constraints_files/figure-html/overlay_perf-1.png){width=816}\n:::\n\n```{.r .cell-code}\nsave_png(\"06-baseline-improved.png\")\n```\n:::\n\n\nComparing all three:\n\n\n::: {.cell}\n\n```{.r .cell-code}\nbackground +\n  bad +\n  improved +\n  baseline +\n  constrained +\n  performance +\n  taller\n```\n\n::: {.cell-output-display}\n![](constraints_files/figure-html/overlay_all-1.png){width=816}\n:::\n\n```{.r .cell-code}\nsave_png(\"07-baseline-constrained-improved.png\")\n```\n:::\n\n\n# Growth of Controls\n\nVisualize an example of the growth of controls using the Cyentia/RiskRecon State of Third-Party Risk\nManagement 2020 and 2024 reports (data from 2023).\n\nSource:\n\n- 2020 data: <https://www.riskrecon.com/state-of-third-party-risk-management-report>\n- 2023 data: <https://www.riskrecon.com/state-of-third-party-risk-management-2024>\n\n\n::: {.cell}\n\n```{.r .cell-code}\nquestionnaire <- tribble(\n  ~year, ~questions, ~percent,\n  2020,  \">400\",     0.041,\n  2020,  \"101-400\",  0.148,\n  2020,  \"11-100\",   0.705,\n  2020,  \"1-10\",     0.107,\n  2023,  \">400\",     0.02,\n  2023,  \"101-400\",  0.333,\n  2023,  \"11-100\",   0.616,\n  2023,  \"1-10\",     0.030\n) |>\n  mutate(year = as.factor(year)) |>\n  mutate(questions = factor(questions, levels = c(\"1-10\", \"11-100\", \"101-400\", \">400\")))\n\nggplot(questionnaire, aes(questions, percent, fill = year)) +\n  geom_col(position = \"dodge\") +\n  scale_y_continuous(labels = scales::label_percent()) +\n  scale_fill_manual(values = c(\"steelblue2\", \"steelblue4\")) +\n  labs(x = NULL, y = NULL, fill = \"Year\", title = \"Third-Party Questionnaire Length by Year\") +\n  labs(caption = \"Source: Cyentia/RiskRecon State of Third-Party Risk Management, 2020 and 2024\") +\n  theme_quo()\n```\n\n::: {.cell-output-display}\n![](constraints_files/figure-html/questionnaire-1.png){width=816}\n:::\n\n```{.r .cell-code}\nsave_png(\"08-questionnaire-length.png\")\n```\n:::\n",
+    "markdown": "---\ntitle: \"Constraints vs Performance\"\nauthor: \"John Benninghoff\"\ndate: '2024-04-30'\ndate-modified: '2024-04-30'\ncategories: notes\norder: 107\noutput:\n  html_notebook:\n    theme:\n      version: 5\n      preset: bootstrap\n    css: assets/extra.css\n    pandoc_args: --shift-heading-level-by=1\n    toc: yes\n    toc_float:\n      collapsed: no\n      smooth_scroll: no\n---\n\n\nVisualizations exploring the use of constraints vs performance improvements in risk management.\n\n# Questions/TODO\n\n- [ ] Questions/TODO list here\n\n\n::: {.cell}\n\n```{.r .cell-code}\nlibrary(ggplot2)\nlibrary(jbplot)\nlibrary(fs)\nlibrary(tibble)\nlibrary(dplyr)\n```\n:::\n\n\n# Normal Performance\n\nReplicate a version of Figure 9 from the\n[Safety-II White Paper](https://www.england.nhs.uk/signuptosafety/wp-content/uploads/sites/16/2015/10/safety-1-safety-2-whte-papr.pdf),\nwith help from <https://ggplot2tutor.com/tutorials/sampling_distributions>:\n\n\n::: {.cell}\n\n```{.r .cell-code}\nxmin <- -5\nxmax <- 5\n\nsave_png <- function(filename) {\n  ggsave(filename = path(\"rendered\", filename), width = 16 * 0.6, height = 9 * 0.6, bg = \"white\")\n}\n\nbackground <- ggplot(data.frame(x = c(xmin, xmax)), aes(x)) +\n  scale_x_continuous(breaks = -3:3, minor_breaks = NULL) +\n  labs(x = NULL, y = NULL) +\n  theme_quo(minor.y = FALSE)\n\nbaseline <- stat_function(fun = dnorm, geom = \"line\")\nbad <- stat_function(fun = dnorm, geom = \"area\", fill = \"red\", xlim = c(xmin, -2))\n\nbackground + bad + baseline\n```\n\n::: {.cell-output-display}\n![](constraints_files/figure-html/background-1.png){width=816}\n:::\n\n```{.r .cell-code}\nsave_png(\"01-baseline-bad.png\")\n```\n:::\n\n\nThe plot above shows \"bad\" outcomes in red. Let's add in \"good\" outcomes (>1) in green:\n\n\n::: {.cell}\n\n```{.r .cell-code}\ngood <- stat_function(fun = dnorm, geom = \"area\", fill = \"green\", xlim = c(1, xmax))\n\nbackground + bad + good + baseline\n```\n\n::: {.cell-output-display}\n![](constraints_files/figure-html/good-1.png){width=816}\n:::\n\n```{.r .cell-code}\nsave_png(\"02-baseline-bad-good.png\")\n```\n:::\n\n\n# Constrained Performance\n\nOne way of reducing \"bad\" outcomes is by constraining performance - reducing the standard\ndeviation.\n\n\n::: {.cell}\n\n```{.r .cell-code}\nconstrained <- stat_function(fun = dnorm, args = list(sd = 0.7), geom = \"line\", color = \"blue\")\ntaller <- scale_y_continuous(limits = c(0, 0.6))\n\nbackground +\n  stat_function(\n    fun = dnorm, args = list(sd = 0.7), geom = \"area\", fill = \"red\", xlim = c(xmin, -2)\n  ) +\n  stat_function(\n    fun = dnorm, args = list(sd = 0.7), geom = \"area\", fill = \"green\", xlim = c(1, xmax)\n  ) +\n  constrained +\n  taller\n```\n\n::: {.cell-output-display}\n![](constraints_files/figure-html/constrained-1.png){width=816}\n:::\n\n```{.r .cell-code}\nsave_png(\"03-constrained.png\")\n```\n:::\n\n\nPlotting both on the same grid shows the reduction in both \"bad\" and \"good\" outcomes:\n\n\n::: {.cell}\n\n```{.r .cell-code}\nbackground +\n  bad +\n  good +\n  baseline +\n  constrained +\n  taller\n```\n\n::: {.cell-output-display}\n![](constraints_files/figure-html/overlay-1.png){width=816}\n:::\n\n```{.r .cell-code}\nsave_png(\"04-baseline-constrained.png\")\n```\n:::\n\n\n# Improved Performance\n\nAnother way of reducing bad outcomes is by improving performance - shifting the mean.\n\n\n::: {.cell}\n\n```{.r .cell-code}\nperformance <- stat_function(fun = dnorm, args = list(mean = 1), geom = \"line\", color = \"blue\")\nimproved <- stat_function(\n  fun = dnorm, args = list(mean = 1), geom = \"area\", fill = \"green\", xlim = c(1, xmax)\n)\n\nbackground +\n  stat_function(\n    fun = dnorm, args = list(mean = 1), geom = \"area\", fill = \"red\", xlim = c(xmin, -2)\n  ) +\n  improved +\n  performance\n```\n\n::: {.cell-output-display}\n![](constraints_files/figure-html/performance-1.png){width=816}\n:::\n\n```{.r .cell-code}\nsave_png(\"05-improved.png\")\n```\n:::\n\n\nPlotting both together shows a reduction in \"bad\" and an increase in \"good\" outcomes:\n\n\n::: {.cell}\n\n```{.r .cell-code}\nbackground +\n  bad +\n  improved +\n  baseline +\n  performance\n```\n\n::: {.cell-output-display}\n![](constraints_files/figure-html/overlay_perf-1.png){width=816}\n:::\n\n```{.r .cell-code}\nsave_png(\"06-baseline-improved.png\")\n```\n:::\n\n\nComparing all three:\n\n\n::: {.cell}\n\n```{.r .cell-code}\nbackground +\n  bad +\n  improved +\n  baseline +\n  constrained +\n  performance +\n  taller\n```\n\n::: {.cell-output-display}\n![](constraints_files/figure-html/overlay_all-1.png){width=816}\n:::\n\n```{.r .cell-code}\nsave_png(\"07-baseline-constrained-improved.png\")\n```\n:::\n\n\n# Growth of Controls\n\nVisualize an example of the growth of controls using the Cyentia/RiskRecon State of Third-Party Risk\nManagement 2020 and 2024 reports (data from 2023).\n\nSource:\n\n- 2020 data: <https://www.riskrecon.com/state-of-third-party-risk-management-report>\n- 2023 data: <https://www.riskrecon.com/state-of-third-party-risk-management-2024>\n\n\n::: {.cell}\n\n```{.r .cell-code}\nquestionnaire <- tribble(\n  ~year, ~questions, ~percent,\n  2020,  \">400\",     0.041,\n  2020,  \"101-400\",  0.148,\n  2020,  \"11-100\",   0.705,\n  2020,  \"1-10\",     0.107,\n  2023,  \">400\",     0.02,\n  2023,  \"101-400\",  0.333,\n  2023,  \"11-100\",   0.616,\n  2023,  \"1-10\",     0.030\n) |>\n  mutate(year = as.factor(year)) |>\n  mutate(questions = factor(questions, levels = c(\"1-10\", \"11-100\", \"101-400\", \">400\")))\n\nggplot(questionnaire, aes(questions, percent, fill = year)) +\n  geom_col(position = \"dodge\") +\n  scale_y_continuous(labels = scales::label_percent()) +\n  scale_fill_manual(values = c(\"steelblue2\", \"steelblue4\")) +\n  labs(x = NULL, y = NULL, fill = \"Year\", title = \"Third-Party Questionnaire Length by Year\") +\n  labs(caption = \"Source: Cyentia/RiskRecon State of Third-Party Risk Management, 2020 and 2024\") +\n  theme_quo()\n```\n\n::: {.cell-output-display}\n![](constraints_files/figure-html/questionnaire-1.png){width=816}\n:::\n\n```{.r .cell-code}\nsave_png(\"08-questionnaire-length.png\")\n```\n:::\n\n\n# Transparent Donut\n\nCreate a transparent donut plot showing an 80% reduction.\n\n\n::: {.cell}\n\n```{.r .cell-code}\n# custom function based on ggplot_donut_percent()\ncustom_donut <- function(p, text = \"\", accuracy = NULL, hsize = 4, size = 12, family = \"Lato\") {\n  data <- data.frame(group = c(TRUE, FALSE), n = c(p, 1 - p))\n  label <- paste0(scales::label_percent(accuracy = accuracy)(p), \"\\n\", text)\n\n  ggplot_donut(data, hsize = hsize) +\n    guides(fill = \"none\") +\n    geom_text(x = 0, label = label, size = size, family = family) +\n    scale_fill_grey() +\n    theme(plot.background = element_rect(fill = \"transparent\", color = NA))\n}\n\ncustom_donut(0.8, \"reduction\")\n```\n\n::: {.cell-output-display}\n![](constraints_files/figure-html/custom_donut-1.png){width=816}\n:::\n\n```{.r .cell-code}\nggsave(\"rendered/80-percent-safety.png\")\n```\n\n::: {.cell-output .cell-output-stderr}\n\n```\nSaving 8.5 x 5 in image\n```\n\n\n:::\n\n```{.r .cell-code}\ncustom_donut(0.8, \"reduction?\")\n```\n\n::: {.cell-output-display}\n![](constraints_files/figure-html/custom_donut-2.png){width=816}\n:::\n\n```{.r .cell-code}\nggsave(\"rendered/80-percent-security.png\")\n```\n\n::: {.cell-output .cell-output-stderr}\n\n```\nSaving 8.5 x 5 in image\n```\n\n\n:::\n:::\n",
     "supporting": [
       "constraints_files"
     ],

analysis/constraints.Rmd

@@ -185,3 +185,27 @@ ggplot(questionnaire, aes(questions, percent, fill = year)) +
 
 save_png("08-questionnaire-length.png")
 ```
+
+# Transparent Donut
+
+Create a transparent donut plot showing an 80% reduction.
+
+```{r custom_donut}
+# custom function based on ggplot_donut_percent()
+custom_donut <- function(p, text = "", accuracy = NULL, hsize = 4, size = 12, family = "Lato") {
+  data <- data.frame(group = c(TRUE, FALSE), n = c(p, 1 - p))
+  label <- paste0(scales::label_percent(accuracy = accuracy)(p), "\n", text)
+
+  ggplot_donut(data, hsize = hsize) +
+    guides(fill = "none") +
+    geom_text(x = 0, label = label, size = size, family = family) +
+    scale_fill_grey() +
+    theme(plot.background = element_rect(fill = "transparent", color = NA))
+}
+
+custom_donut(0.8, "reduction")
+ggsave("rendered/80-percent-safety.png")
+
+custom_donut(0.8, "reduction?")
+ggsave("rendered/80-percent-security.png")
+```

docs/NEWS.html

@@ -235,7 +235,8 @@
     <h2 id="toc-title">On this page</h2>
 
   <ul>
-  <li><a href="#rtraining-1.2.4" id="toc-rtraining-1.2.4" class="nav-link active" data-scroll-target="#rtraining-1.2.4">rtraining 1.2.4</a></li>
+  <li><a href="#rtraining-1.2.5" id="toc-rtraining-1.2.5" class="nav-link active" data-scroll-target="#rtraining-1.2.5">rtraining 1.2.5</a></li>
+  <li><a href="#rtraining-1.2.4" id="toc-rtraining-1.2.4" class="nav-link" data-scroll-target="#rtraining-1.2.4">rtraining 1.2.4</a></li>
   <li><a href="#rtraining-1.2.3" id="toc-rtraining-1.2.3" class="nav-link" data-scroll-target="#rtraining-1.2.3">rtraining 1.2.3</a></li>
   <li><a href="#rtraining-1.2.2" id="toc-rtraining-1.2.2" class="nav-link" data-scroll-target="#rtraining-1.2.2">rtraining 1.2.2</a></li>
   <li><a href="#rtraining-1.2.1" id="toc-rtraining-1.2.1" class="nav-link" data-scroll-target="#rtraining-1.2.1">rtraining 1.2.1</a></li>
@@ -325,6 +326,12 @@ <h2 id="toc-title">On this page</h2>
 
 
 
+<section id="rtraining-1.2.5" class="level1">
+<h1>rtraining 1.2.5</h1>
+<ul>
+<li>Updated Constraints vs Performance: Added “Transparent Donut” section</li>
+</ul>
+</section>
 <section id="rtraining-1.2.4" class="level1">
 <h1>rtraining 1.2.4</h1>
 <ul>