From d7ddf39326eda15785ad58b6ffc2c496b0ce31db Mon Sep 17 00:00:00 2001
From: Jose <greenleaves@disroot.org>
Date: Tue, 25 Oct 2022 20:32:51 -0300
Subject: [PATCH] Cleaning the data

---
 script/cleaning.R | 108 ++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 108 insertions(+)
 create mode 100644 script/cleaning.R

diff --git a/script/cleaning.R b/script/cleaning.R
new file mode 100644
index 0000000..710a7ac
--- /dev/null
+++ b/script/cleaning.R
@@ -0,0 +1,108 @@
+#' ---
+#' title: Emacs survey 2020
+#' date: "2022-10-25"
+#' author: "Jose - https://ajuda.multifarm.top"
+#' output:
+#'   html_document:
+#'    code_folding: show
+#'    toc: yes
+#'    toc_float:
+#'      smooth_scroll: true
+#'    df_print: paged
+#'    highlight: zenburn
+#' ---
+
+rm(list = ls())
+
+#   renv::init() #' For reproducibility
+
+library(data.table)
+library(DT)
+
+#' ## Emacs survey 2020:
+#'
+#' cleaned up data:
+#'
+#' https://emacs-survey.netlify.app/2020/Emacs-User-Survey-2020-clean.csv
+#'
+#' Raw data:
+#'
+#' https://emacs-survey.netlify.app/2020/Emacs-User-Survey-2020-raw.csv
+
+url = 'https://emacs-survey.netlify.app/2020/Emacs-User-Survey-2020-clean.csv'
+
+d <- fread(url, check.names = TRUE)
+
+str(d)
+
+names(d)
+
+#' ## Date
+#'
+#' Should arrange
+
+
+class(d$Submission.Date)
+
+d[, .N, Submission.Date]
+
+#' ## Favorite packages
+
+DT::datatable(d[, .N, Can.you.list.some.of.your.favorite.packages.][order(-N)])
+
+#' ## Difficulties faced
+
+d[, .N, Can.you.recall.any.difficulties.you.faced.initially.learning.Emacs.][
+    order(-N)] |>
+    datatable()
+
+
+#' ## Describe org-mode usage
+
+datatable(d[, .N, Describe.your.org.mode.usage])
+
+
+#' ## Describe org-mode usage
+
+d[, .N, For.how.many.years.have.you.been.using.Emacs.][order(-N)] |>
+    knitr::kable()
+
+sent1 =  'I installed it first circa 2013, I guess.  Hard to remember; it was on Windows.  I was not a regular user until 2016, and a confident user until 2017.'
+
+sent2 = "I've used Emacs for nearly one decade, but will be purposefully
+vague beyond this."
+
+sent1
+
+sent2
+
+d[For.how.many.years.have.you.been.using.Emacs. == sent1,
+  "For.how.many.years.have.you.been.using.Emacs."] <- 7
+
+d[For.how.many.years.have.you.been.using.Emacs. == sent2,
+  "For.how.many.years.have.you.been.using.Emacs."] <- 10
+
+
+d[, .N, For.how.many.years.have.you.been.using.Emacs.]
+
+d[, Years.using.emacs :=
+        as.numeric(For.how.many.years.have.you.been.using.Emacs.)]
+
+d[, .N, Years.using.emacs]
+
+library(ggplot2)
+
+ggplot(d, aes(Years.using.emacs)) +
+    geom_histogram() +
+    theme_minimal()
+
+ggplot(d, aes(Years.using.emacs)) +
+    geom_density(aes(color = "red", fill = "red"), alpha = 0.3) +
+#    geom_histogram(alpha = 0.1) +
+    theme_minimal()
+
+#' ## OS you primarily use with emacs
+
+datatable(d[, .N, Which.OS.do.you.primarily.use.Emacs.on.][
+    order(-N)
+])