Generate Bingo cards with R.
R CSS
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
R
img
inst
man
tests
.Rbuildignore
.gitignore
CONTRIBUTING.md
DESCRIPTION
LICENSE
NAMESPACE
README.Rmd
README.md
bingo.Rproj

README.md

bingo

Generate Bingo cards.

Currently has built-in squares for SuperBowl 50 🏈 and data / spreadsheet craziness 📉. Or you can provide your own text for the squares.

Make printable Bingo cards without installing anything via this Shiny app:

Feel free to help us make these cards less ugly or to explore new bingo topics! PRs welcome 😁.

Installation

Install from GitHub with:

# install.packages("devtools")
devtools::install_github("jennybc/bingo")

SuperBowl Example

library(bingo)

## see some of the SuperBowl 50 squares
tail(get_topic("football"))
#> [1] "Shot of Golden Gate Bridge"                
#> [2] "\"Silicon Valley and tech\" blah blah blah"
#> [3] "Mike Carey is WRONG"                       
#> [4] "Unexpected artist joins Beyoncé"           
#> [5] "Cam's Superman shirt-opening thing"        
#> [6] "Idle speculation it's Peyton's last game"

## make 8 bingo cards (SuperBowl 50 is current default)
bc <- bingo(8)

## print them to PDF
plot(bc)
#> Writing to file ...
#>   ./bingo-01.pdf
#>   ./bingo-02.pdf
#>   ./bingo-03.pdf
#>   ./bingo-04.pdf
#>   ./bingo-05.pdf
#>   ./bingo-06.pdf
#>   ./bingo-07.pdf
#>   ./bingo-08.pdf

Here's what one looks like:

"Open" and Bad Data Examples

We offer two sets of squares inspired by the pain joy of dealing with #otherpeoplesdata

Use get_topic("open-data") to get squares based on this tweet from Chris McDowall:

For two weeks I noted issues encountered as I used NZ govt data. Today I collected enough to make a bingo card. @fogonwater, January 3, 2016

Use get_topic("bad-data") to get squares inspired by the Quartz guide to bad data:

An exhaustive reference to problems seen in real-world data along with suggestions on how to resolve them.... Most of these problems can be solved. Some of them can't be solved and that means you should not use the data. Others can't be solved, but with precautions you can continue using the data.

## see some Open Data squares
tail(get_topic("open-data"))
#> [1] "PDF tables"                       "numbers formatted as text"       
#> [3] "metadata expressed as fieldnames" "fieldname EDA"                   
#> [5] "regex-driven workflow"            "named region non-sequiturs"

## see some Bad Data squares
tail(get_topic("bad-data"))
#> [1] "Data disguised as formatting"                  
#> [2] "Ambiguous American date formats, eg 03/04/16"  
#> [3] "\"Virgin Birth\", ie no provenance"            
#> [4] "Location of 0°N 0°E, ie \"Null Island\""       
#> [5] "Spelling mistakes that reek of hand-typed data"
#> [6] "US zip codes 12345 or 90210"

## make a single Open Data bingo card
bc <- bingo(words = get_topic("open-data"))

## make a custom bingo blend from the open and bad data squares
bc <- bingo(words = c(get_topic("open-data"), get_topic("bad-data")))

## print it
plot(bc, pdf_base = "open-data-")
#> Writing to file ...
#>   ./open-data-01.pdf

Here's an Open Data bingo card:

Run Shiny app locally

To run the app we're running remotely on your own machine, do this:

launch()