Baby Name MaxDiff

A Shiny app that ranks baby names for two parents using MaxDiff.

The app's results screen. It recommends Spencer Curtis and shows a bar chart of utility scores for nine names, one bar per spouse, with a red mark at the average. Spencer and Jackson score highest; Jared and Ryan score lowest.
The results screen. Each name gets a utility score from each spouse, and the red mark is the average of the two.
Try the app Code on GitHub

At Qualtrics one of my team's primary revenue drivers was MaxDiff studies. MaxDiff is a tool that makes it easier to rank long lists of items by asking respondents to choose a favorite and least favorite from a small subset of all the choices. My team joked about using MaxDiff for all kinds of personal decisions, but especially baby names. Naturally, when my wife and I got pregnant with our first kid in 2021, I decided to make the joke a reality and build a Shiny app that would let you input a bunch of names (or randomly select popular names), take turns picking favorites, and then get recommendations based on what you and your spouse mutually liked.

The app runs on a free shinyapps.io account. It can take a little while to start, and it may be down if the month's free hours have run out. Fitting the Bayesian model to estimate utilities at the end takes a minute or two.

The GitHub repo is an R package. The interface and question logic are in inst/server.R, and the model fit is in R/helper_functions.R. To run it locally, clone the repo and run app.R.