matt-dray / tidyquiz Goto Github PK

View Code? Open in Web Editor NEW

2.0 2.0 2.0 11.43 MB

:package::question: R Shiny app: a {tidyverse} quiz via {learnr} hacking

License: Other

R 0.49% HTML 99.51%

r quiz learnr tidyverse shiny rstats

tidyquiz's Introduction

tidyquiz

Can you match the tidyverse function to the package name? A multiple choice quiz hacked together with {learnr}.

What?

Can you always remember which tidyverse function comes from which tidyverse package? Inspired by frustrations of Ryan Timpe, Daniel McNichol, and others I'm sure.

The {learnr} package lets you build quizzes, obstensibly for learning about R. I wondered if you could use it to serve a randomised question about what function belongs to which tidyverse package. It's a bit hacky: you have to click the 'get question' button to begin and then 'start over' each time you want a new question.

You can read the accompanying blog post for more info on how it was built.

How to run

Install the package and run the app with the following commands:

install.packages(c("remotes", "pacman", "tidyverse", "learnr")) to install {learnr} from CRAN (beware: this will install a lot of packages)
remotes::install_github("matt-dray/tidyquiz") to install this package from GitHub
learnr::run_tutorial("tidy", package = "tidyquiz") to open in your browser

The app is no longer hosted online.

Notes

Do I have to install loads of packages for this to work?

Yes, if installing locally. This thing works on the fly. It fetches the tidyverse packages and creates questions and answers whenever you run generate_quiz(). This means you should be prepared for all the tidyverse packages to be downloaded to your machine. On the plus side, functions presented to you in the quiz will always be current.

The list of packages can be found in tidyverse::tidyverse_packages(). At time of writing (2019-03-17) the packages are:

"broom"     "cli"         "crayon"   "dplyr"     "dbplyr"      "forcats"
"ggplot2"   "haven"       "hms"      "httr"      "jsonlite"    "lubridate"
"magrittr"  "modelr"      "purrr"    "readr"     "readxl"      "reprex"
"rlang"     "rstudioapi"  "rvest"    "stringr"   "tibble"      "tidyr"      
"xml2"      "tidyverse"

What about functions that mask other functions?

There's only ever one correct answer from the four presented. If you are presented with a function that exists in multiple packages, the answer set will include one package it definitely exists in and three it definitely doesn't exist in.

tidyquiz's People

Contributors

Stargazers

Watchers

Forkers

weijia2013 korterling

tidyquiz's Issues

Provide overall score

You got 8/10

Can we capture the output that tells users they're successful or not and sum them for an overall score?

Is there an equivalent of shiny::runApp(display.mode="showcase") that might expose this?

Unanswered question here: https://community.rstudio.com/t/learnr-track-student-performance-on-multiple-choice-questions/17458

Try alternative to {learnr}

{shiny}: button to randomise questions; answer with radio buttons/check boxes and 'submit' button; final score tallied.

How about {swirlify}? Can users get a randomised list of questions?

Startup message doesn't appear on load

library(tidyverse)
#> Run `learnr::run_tutorial("tidy", package = "tidyquiz")` to start the quiz

Related to zzz.R. Maybe just needs rebuilding?

.onAttach <- function(libname, pkgname) {
  packageStartupMessage(
    'Run `learnr::run_tutorial("tidy", package = "tidyquiz")` to start the quiz'
  )
}

Deal with conflicts

Ignore conflicts completely (easy, not as exciting)
Allow more than one correct answer

The function StatBinhex is from which tidyverse package?

{ggplot2}

{broom}

{ggplot2}

{xml2}

Related to the method of excluding the correct answer, i.e. from the Rmd:

# Get wrong answer 1 of 3
ans_wrong1 <- eventReactive(
  input$goButton,
  { 
    set.seed(input$n)
    fun_sample <- sample_n(tidy_funs, 1)
    fun_name <- select(fun_sample, functions) %>% pull()
    # PROBLEM HERE?    
    dup_lookup <- tidy_funs %>%  
      filter(functions == fun_name) %>% 
      pull(package)
    ans_correct <- fun_sample %>% select(package) %>% pull()
    ans_wrong1 <- tidy_funs %>%
      distinct(package) %>%  # get unique packages
      # OR MAYBE HERE?
      filter(!package %in% c(dup_lookup)) %>%  # ignore packages containing function
      sample_n(1) %>%  # choose a remaining package name at random
      pull()  # to character
    return(ans_wrong1)
  }
)
output$ans_wrong1_out <- renderText({ paste0("{", ans_wrong1(), "}") })

Refresh the questions

User clicks a button and is presented with a fresh set of questions.

The function a user is presented with in a question is selected randomly, as are the possible packages in the answer selections. Could a button generate and present new samples?