Systemic Handicap Parking Violations
Documented illegal use of handicap parking at Magnolia Academy Children's Center. I infrequently do pickup/dropoff. This is what I've observed.
7300 Magnolia Market Ave, Moseley, VA 23120 · (804) 203-5191 · magnoliaacademyva.com
How this analysis works
Instead of asking "could this be a coincidence?" — the classical question — Bayesian analysis asks a more intuitive one: given what I observed, what should I now believe about how often violations actually happen?
The math takes two inputs and produces one output. Inputs: a starting belief (before seeing any data, every possible violation rate from 0% to 100% is treated as equally plausible — no assumptions baked in) and the data itself (each visit is one observation: violation or no-violation). Output: the curve below — the posterior — showing which violation rates are most consistent with what has been observed.
The width of the curve reflects how much uncertainty remains. With only a handful of observations the curve is wide; with hundreds it tightens. The shaded red region is the portion of belief lying above the 5% baseline — the threshold at which "this is just random chance" stops being a defensible explanation. When most of the curve sits to the right of 5%, the data is incompatible with a random-and-rare model.
The chart shows the posterior estimate of the true violation rate (green curve), the 5% baseline (dashed red line), and the proportion of belief above 5% (shaded red). See the methodology page for the full mathematical detail — Beta-Binomial conjugate model, Bayes factor derivation, and frequentist confirmation.
Why this isn't coincidence
A common objection: "You only visit roughly 10% of pickup/dropoff events. Maybe violations are rare and you just happened to catch them." The math handles this directly.
- Visit rate and violation rate are independent. How often I personally show up doesn't change how often violations happen at any given moment. Each visit is one independent observation of the parking lot. With 8 such observations and 6 violations, the Bayesian model can already pin down the underlying rate with confidence — small samples are enough when the imbalance is this extreme.
- A truly rare violation rate would make this nearly impossible. If violations happened only 5% of the time (the threshold below which "this is just noise" stops being defensible), the chance of seeing 6 violations in 8 random visits is less than one in a million. The data is incompatible with any "rare" model.
- My visits aren't timed around the parking lot. I show up when work and family logistics dictate, not based on what I expect to see (or not see). That matters: random sampling of moments produces an unbiased estimate of the true rate. If anything, accessibility users who arrive at routine times — like I do — are exactly the people the analysis applies to.
- Even a single frequent offender is a systemic problem. Whether the same driver violates every day or different drivers each time, the spaces are unavailable to people who need them. The harm — and the ADA Title III requirement — is that accessible parking actually be accessible.
Recent Incidents
Each visit is logged with date, time of day, and photographic evidence when available. View full log →