In best-practice-and-impact/CARS2: Coding in Analysis and Research Survey (CARS) Wave 2 Analysis

# Set echo=FALSE for all chunks
knitr::opts_chunk$set(echo=FALSE)

Coding frequency and tools

Coding frequency

We asked our respondents, "In my current role, I write code to complete my work objectives _". Most of respondents either coded regularly or all the time.

library(magrittr)

if(!exists("data")) stop("Dataframe called data not available. This should be in the 
                                      function enviroment of render_main_site. Check that this is available in this
                                      enviroment.")

# Create tables
tables <- carsurvey2::generate_tables(data)

plot <- carsurvey2::plot_freqs(tables$freq_table, "Coding frequency", "Count", n = samples$all, font_size = 14, width = 600)

table <- kableExtra::kable_styling(knitr::kable(tables$freq_table)) %>% 
  kableExtra::add_footnote(paste0("Sample size = ", samples$all))

carsurvey2::wrap_outputs("coding-freq", plot, table)

Access to and knowledge of programming languages

For each of the most popular programming languages from last year's CARS data, we asked respondents to answer "yes", "no" or "I don't know" for the following statements:

• I know how to program with this tool to a level suitable for my work
• This tool is available to use for my work

Knowledge of programming tools

Most common programming tool was R followed by SQL.

plot <- carsurvey2::plot_stacked(tables$knowledge, xlab = "Count", ylab = "Programming tool", n = samples$all, colour_scale = "3scale", font_size = 14, width = 600)

table <- kableExtra::kable_styling(knitr::kable(tables$knowledge, row.names = FALSE)) %>%  kableExtra::add_footnote(paste0("Sample size = ", samples$all))

carsurvey2::wrap_outputs("tools-knowledge", plot, table)

Access to programming tools

The programming tools with the greatest access were R and Python respectively.

plot <- carsurvey2::plot_stacked(tables$access, xlab = "Count", ylab = "Programming tool", n = samples$all, colour_scale = "3scale", font_size = 14, width = 600)

table <- kableExtra::kable_styling(knitr::kable(tables$access, row.names = FALSE)) %>%
  kableExtra::add_footnote(paste0("Sample size = ", samples$all))

carsurvey2::wrap_outputs("tools-access", plot, table)

Access and knowledge gaps

We used the above data to calculate the number of respondents who have access but no knowledge, access and knowledge, and knowledge but no access for each programming language. The tool with the greatest access only was Python and the tool with the greatest access and knowledge was R.

plot <- carsurvey2::plot_stacked(tables$code_tool_status, colour_scale = "3scale", xlab = "Count", ylab = "Programming tool", n = samples$all, font_size = 14, width = 600)

table <- kableExtra::kable_styling(knitr::kable(tables$code_tool_status, row.names = FALSE)) %>%
  kableExtra::add_footnote(paste0("Sample size = ", samples$all))

carsurvey2::wrap_outputs("tools-status", plot, table)

What are people using code for?

We asked respondents what data operations they do in their work, and whether they use code to do them. The data operation with the highest frequency done with some or all code is data analysis. The data operation with the highest frequency done without coding is quality assurance.

plot <- carsurvey2::plot_stacked(tables$freq_ops, xlab = "Count", ylab = "Data operation", n = samples$all, neutral_mid = FALSE, font_size = 14, width = 600)

table <- kableExtra::kable_styling(knitr::kable(tables$freq_ops, row.names = FALSE)) %>% 
  kableExtra::add_footnote(paste0("Sample size = ", samples$all))

carsurvey2::wrap_outputs("operations", plot, table)

Coding capability

We asked respondents a series of questions about whether they feel their coding ability is improving in their current role and where they first learned to code.

Change in coding ability during current role

Respondents who had coding experience outside of their current role were asked whether there has been a change in their coding ability during current employment. Most respondents coding ability improved during current role.

plot <- carsurvey2::plot_freqs(tables$freq_abil, xlab = "Change in ability", ylab = "Count", n = samples$code_outside_current_role, break_q_names_col = "Coding ability changes", font_size = 14, width = 600)

table <- kableExtra::kable_styling(knitr::kable(tables$freq_abil, row.names = FALSE)) %>% 
  kableExtra::add_footnote(paste0("Sample size = ", samples$code_outside_current_role))

carsurvey2::wrap_outputs("ability", plot, table)

Where respondents have learned to code

Respondents were asked whether they had coding experience outside their current role and, if so, where. Most of the respondents first learnt code in education.

plot <- carsurvey2::plot_freqs(tables$first_learnt, xlab = "", ylab = "Count", n = samples$any_code_experience, break_q_names_col = "First coding experience", orientation = "h", font_size = 14, width = 600)

table <- kableExtra::kable_styling(knitr::kable(tables$first_learnt, row.names = FALSE)) %>% 
  kableExtra::add_footnote(paste0("Sample size = ", samples$any_code_experience))

carsurvey2::wrap_outputs("first-learn", plot, table)

Coding practices

We asked respondents to report how often they make use of each of the coding practices presented below. Please note that while there are many different coding practices listed below, we understand that not all are proportionate for every coding project.

General coding practices

The most used coding practice is “I write repetitive elements in my code as functions”.

plot <- carsurvey2::plot_likert(tables$code_prac_chart, 5, "Frequency", "", n = samples$coders, neutral_mid = FALSE, font_size = 14, height = 700, width = 700, break_q_names_col = "Question")

table <- kableExtra::kable_styling(carsurvey2::output_likert_table(tables$code_prac_chart)) %>% 
  kableExtra::add_footnote(paste0("Sample size = ", samples$coders))

carsurvey2::wrap_outputs("code-prac", plot, table)

Documentation

The most used documentation is code comments.

plot <- carsurvey2::plot_likert(tables$doc, 5, "Frequency", "", n = samples$coders, neutral_mid = FALSE, font_size = 14, height = 700, width = 700, break_q_names_col = "Question")

table <- kableExtra::kable_styling(carsurvey2::output_likert_table(tables$doc)) %>% 
  kableExtra::add_footnote(paste0("Sample size = ", samples$coders))

carsurvey2::wrap_outputs("doc", plot, table)

Reproducible workflow packages

We asked respondents "do you use reproducible workflow packages e.g. drake, make or pymake?". Most respondents didn’t use reproducible workflow packages.

plot <- carsurvey2::plot_freqs(tables$rep_workflow, xlab = "Use reproducible workflows packages", ylab = "count", n = samples$coders, break_q_names_col = "Use reproducible workflow packages", font_size = 14, width = 600)

table <- kableExtra::kable_styling(knitr::kable(tables$rep_workflow)) %>% 
  kableExtra::add_footnote(paste0("Sample size = ", samples$coders))

wrap_outputs("workflow", plot, table)

Source control platform

The number of users of each source control platform. If a source control platform was used most respondents used GitHub.

plot <- carsurvey2::plot_freqs(tables$vers_cont, xlab = "", ylab = "Count", n = samples$coders, orientation = "h", break_q_names_col = "Version control platform", font_size = 14, width = 600)

table <- kableExtra::kable_styling(knitr::kable(tables$vers_cont)) %>% 
  kableExtra::add_footnote(paste0("Sample size = ", samples$coders))

wrap_outputs("version-control", plot, table)

RAP knowledge and opinions

Knowledge of RAP and RAP champions

We asked our respondents whether they had heard of RAP and what their knowledge is of their own department RAP champion. The most frequent response was “Have not heard of RAP”.

plot <- carsurvey2::plot_freqs(tables$rap_knowledge_chart, "", "Count", n = samples$all, font_size = 14, orientation = "h", break_q_names_col = "RAP champion knowledge", max_lines = 3, width = 600)

table <- kableExtra::kable_styling(knitr::kable(tables$rap_knowledge, row.names = FALSE)) %>% 
  kableExtra::add_footnote(paste0("Sample size = ", samples$all))

carsurvey2::wrap_outputs("rap-knowledge", plot, table)

Opinions on RAP

We asked our respondents who had heard of RAP the extent to which they agree with the following statements:

• "I understand what the key components of the RAP methodology are"
• "I feel confident implementing RAP in my work"
• "I think it is important to implement RAP in my work"
• "I feel supported to implement RAP in my work"
• "I know where to find resources to help me implement RAP"
• "I and/or my team are currently implementing RAP"

The figure and table show the percentage of respondents who picked each response option. Percentages are out of a sample of respondents who said they had heard of RAP. The statement most respondents agree with is “I think it is important to implement RAP in my work”.

plot <- carsurvey2::plot_likert(tables$rap_opinions_chart, 3, "Frequency", "", n = samples$heard_of_rap, font_size = 14, height = 700, width = 700, break_q_names_col = "Question")

table <- kableExtra::kable_styling(carsurvey2::output_likert_table(tables$rap_opinions)) %>% 
  kableExtra::add_footnote(paste0("Sample size = ", samples$heard_of_rap))

carsurvey2::wrap_outputs("rap-opinions", plot, table)

RAP scores

RAP components

The frequencies for each RAP component below are derived from a series of questions on coding practices (see the coding practices page). These were used to calculate scores for each component (0/1) where respondents answered "regularly" or "all the time" to the relevant questions. The list of RAP components used here is based on the RAP minimum viable product guidance agreed by the RAP champions network. The RAP components most used were “Functions” and “Follow code style guidelines” respectively which are both in “Advanced” components.

plot <- carsurvey2::plot_grouped(tables$components, "", "Count", n = samples$coders, orientation = "h", font_size = 14, width = 600)

table <- kableExtra::kable_styling(knitr::kable(tables$components, row.names = FALSE)) %>% 
  kableExtra::add_footnote(paste0("Sample size = ", samples$coders))

carsurvey2::wrap_outputs("rap_comps", plot, table)

Basic RAP scores

bar_colour <- carsurvey2::get_2colour_scale(2)[2][[1]]
bar_colour <- rgb(bar_colour[1],bar_colour[2],bar_colour[3], maxColorValue = 255)

plot <- carsurvey2::plot_freqs(tables$basic_freqs, "Basic RAP score", "Count", bar_colour = bar_colour, n = samples$coders, font_size = 14, width = 600)

table <- kableExtra::kable_styling(knitr::kable(tables$basic_freqs, row.names = FALSE)) %>% 
  kableExtra::add_footnote(paste0("Sample size = ", samples$coders))

carsurvey2::wrap_outputs("basic_score", plot, table)

Advanced RAP scores

plot <- carsurvey2::plot_freqs(tables$advanced_freqs, "Advanced RAP score", "Count", n = samples$coders, font_size = 14, width = 600)

table <- kableExtra::kable_styling(knitr::kable(tables$advanced_freqs, row.names = FALSE)) %>% 
  kableExtra::add_footnote(paste0("Sample size = ", samples$coders))

carsurvey2::wrap_outputs("advanced_score", plot, table)

best-practice-and-impact/CARS2 documentation built on Dec. 19, 2021, 8:47 a.m.