How Seen NC works
Seen NC turns a public data feed the state barely surfaces into something you can actually use before a DMV trip. Here's the approach — the honest version, minus the parts that took the longest to get right.
The data is the state's own
North Carolina publishes a live wait-time feed for its driver-license offices. We read it on a schedule through the day and — this is the part that matters — we keep the history. The state shows you only this instant; we've been archiving every reading, building the longitudinal record of how each office actually performs over time. That archive can't be recreated after the fact, which is what makes it valuable.
We rank on measured reality, not guesses
Each office's quietness comes from its real, measured waits — no appointment scraping, no crowd-sourced hearsay, no made-up numbers. The list you see is ordered by what the data says, and nothing else moves an office up it: not lodging, not sponsors, not who's paying. Exactly how we turn thousands of raw readings into one trustworthy score is the part we keep under the hood — getting it right (and keeping it honest) is most of the work.
We're honest about where a number falls short
The state's basic feed reports a wait for every open office, metros included. At 20 metro and appointment-heavy offices, the overflow moves into a text-in virtual queue once the lobby fills, and people are turned away when it closes. Below capacity the lobby clock is a fair read. At capacity it becomes a floor on the visit rather than a measure of it. So we do two things: we leave those offices out of every lobby-derived figure and ranking, and we pull from the state's newer capacity feed, the one that counts the people waiting in their cars, to show a wait and an hour-by-hour busyness pattern where we can. Knowing what your data can't tell you, then finding the source that can, is most of the work.
What each number on an office page means
Every figure is computed for that office alone, one value per observed day, then the median across days, so a day we polled more often does not count for more.
- Typical wait: the median reported lobby wait while the office was open. The feed's lowest value is 0, so "0 min or less" means at or below that floor on most readings, not a measured empty lobby.
- Typical upper reading: each day's 90th-percentile reading, then the median across days.
- At closing: the last wait the feed reported before the office stopped taking people. The feed zeroes its counter at that moment, so the reading right after is discarded as a reset.
- Typical daily peak: the median of each day's highest reported wait. Worst recorded: the highest single reading in the archive, after discarding any wait longer than the office had been open.
- Calmness rank: the office's place among measured offices by the same score the office list is sorted by. 1st is the calmest; equal scores share a rank.
- Counter reset before close: the feed zeroes an office's wait counter when it stops accepting new arrivals, while the people already in line are still served. This is the median time that happened, across the observed closings on which it did. We read the feed every 10 minutes, so times are accurate to about 10 minutes.
- Two closing times: the state's location directory and its live feed often list different closing times. Where they differ, the page says how often the feed still showed the office open at the directory's time.
- Month by month: the same daily figures, combined per calendar month, for months with at least 8 observed days.
- Charts: median wait per open hour and per weekday at that office. The weekday chart appears once every weekday has 3 observed days. Estimates, never a promised wait.
Fast, mobile-first, and disposable
It's built for one hand on a hot-car connection: a lightweight app that opens instantly, refreshes itself while you sit deciding whether to wait or drive, auto-translates to your phone's language, and installs to your home screen if you want it — then deletes clean when your DMV run is done. No account, no login, nothing to remember.
Never a fabricated number
The rule the whole thing is built on: if we don't have real data for something, we say so. A missing number is always an honest "not enough data yet," never a plausible-looking guess. That's harder than it sounds, and it's the difference between a tool you can trust at 8am when it matters and one you can't.
Seen NC is one of ours. We built the record nobody else keeps, then made it useful. If your business is sitting on data it can't get at, that's the work.
clarityforge.dev →