/JSE_OkCupid

Journal of Statistical Education Paper on Using OkCupid Data for Data Science Courses

Primary LanguageR

OkCupid Profile Data for Intro Stats and Data Science Courses

Albert Y. Kim & Adriana Escobedo-Land

Data and code for OkCupid Profile Data for Introductory Statistics and Data Science Courses (Journal of Statistics Education July 2015, Volume 23, Number 2).

  • JSE.bib: bibliography file
  • JSE.pdf: PDF of document
  • JSE.Rnw: R Sweave document to recreate JSE.pdf.
  • JSE.R: R code used in document
  • okcupid_codebook.txt: codebook for all variables
  • profiles.csv.zip: CSV file of profile data (unzip this first)

Note:

  • Permission to use this data set was explicitly granted by OkCupid.
  • Usernames are not included.
  • JSE.Rnw Sweave document was compiled using the knitr package. In RStudio, go to "Tools" -> "Project Options" -> "Sweave" -> "Weave Rnw files using:" and select knitr.

Preview

Distribution of Male and Female Heights

Joint Distribution of Sex and Sexual Orientation

A mosaicplot of the cross-classification of the 59946 users' sex and sexual orientation:

Logistic Regression to Predict Gender

Linear regression (in red) and logistic regression (in blue) compared. Note both the x-axis (height) and y-axis (is female: 1 if user is female, 0 if user is male) have random jitter added to better visualize the number of points involved for each (height x gender) pair.

Fitted probabilities p-hat of each user being female along witha decision threshold (in red) used to predict if user is female or not.