Ranking Names Out Loud

In a classroom, at the park, or at work, we hear names spoken out loud more than we see them spelled out. Pronunciations turn heads, regardless of how you spell a name or what it means.

Whether you're looking for a name that stands out or one that fits in, grouping names that sound alike is key. But official sources only rank spellings, so they miss the big picture, and other sites stop short of grouping every name in the SSA dataset for every year that data is available. This is my attempt to fill in the gaps.

To build this site, I analyzed 103,564 spellings with AI to create the internet's most comprehensive ranking of baby names by pronunciation.

Unlike most AI-influenced content on the internet, significant care and attention went into building something that comes as close as possible to the actual truth. There are certainly still mistakes to be found, but that's because the problem is difficult, not because I'm trying to drown you in BS.

I'm working on a technical blog post explaining my grouping algorithm. The gist of it is that I carefully prompted a generative AI model to come up with pronunciations for each name, developed a custom similarity scoring system for comparing pronunciations, and then used a greedy matching algorithm to group the names based on their pronunciations. In future versions I plan to incorporate user feedback to improve the groupings and provide more visibility into the generated pronunciations.

Your feedback matters! Click on a spelling in the spellings list to vote whether it belongs with that name (and see how others have voted). Name pronunciations are inherently subjective and the models that I'm working with are imperfect. My goal is to have the grouping models update automatically in response to user feedback, although the automated pipelines are still a work in progress.

Click on a name in the rankings to see a chart of all the spellings that are grouped together. I'm working on adding more content to these "combined" name pages, like the pronunciations on which the groupings are based and some more charts of rankings over time. I haven't worked out how I'm going to announce new features yet; I'm bad at social media but am open to suggestions. All I can suggest for now is to check back periodically.

I had a lot of fun building this. I really hope you like it.