This workflow searches through the unicode character set (including emoji) using the Unicode Common Locale Data Repository. This repository has a set of tags and for the characters to ease finding them.
For example, the 🎉 has the text-to-speech name of "party popper" and the tags "ball", "celebrate", "celebration", "confetti", "party" and "woohoo". Any of these things will trigger a match.
In a different plugin I used Levenshtein distance, but the annotations from CLDR a decent enough that just feeding that full set into Alfred's results filter keeps things simple (and really fast since I can rely on Alfred's caching rather than re-running the script every keystroke). It does mean that you have to spell something like ellipsis correct, though, and won't get proper results if you type eli.
Download the newest version via the latest release
There are a few of these that exist, but I didn't quite like how any of them worked. Mostly I'm tired of Googling to get ‽ and ✔, a couple characters that I like to use often.
- Unicode Common Locale Data Repository