Your Neighborhood Is Lying to You

We have been living inside a hallucination. The quiet street, the low-crime zip code, the neighborhood app full of lost-dog posts — it feels safe because the data says safe, and the data says safe because most of what actually happens never makes it into the data. Fifty-two percent of all violent crimes in the United States go unreported to police, according to the Bureau of Justice Statistics National Crime Victimization Survey. More than half. The picture you have built of your block — the mental map you use to decide where to walk at night, where to let your kids go alone, where to buy a house — is assembled from a minority of events. The rest dissolved without a trace.

Hold on. Get comfortable with that number. Because it gets worse.

The Reporting Problem Is Just the First Layer

Sexual assault does not go unreported at 52 percent. It goes unreported at 76 percent. Three out of four sexual assaults in this country leave no police record, no case number, no data point in any database you can access. According to the BJS National Crime Victimization Survey, 76 percent of sexual assaults are never reported to law enforcement. They happened. They just do not exist, officially. The streets where they happened still look like quiet streets on every map you can pull up on your phone.

Now take the crimes that do get reported. The ones that survive that first filter and land in the hands of a law enforcement agency. What happens to them. According to BJS data, 56 percent of reported violent crimes go unsolved. More than half. Reported, investigated, and still — no resolution. No arrest, no closure, no entry in the court system. The event exists as a data point but goes nowhere.

Run those two numbers together. Start with 100 violent crimes. Forty-eight get reported. Of those, roughly 21 result in any kind of clearance. So out of 100 actual violent crimes in your area, you might know about 21 of them — at the absolute outer limit of what the system can tell you. The other 79 are ghosts on your street map.

The Data Holes Go All the Way Down

We are not even done. Because even the official reported data — the floor, the bare minimum representation of reality — is itself riddled with holes. In 2021, only about half of all law enforcement agencies in the United States submitted complete crime data to the FBI's Uniform Crime Reporting program. Half. The FBI, the single authoritative national source on crime statistics, was working with incomplete returns from more than half the agencies it was supposed to be collecting from. The national violent crime picture is built on partial submissions from a patchwork of jurisdictions that did not all show up to hand in their numbers.

The national violent crime rate for 2023 was 369.8 per 100,000 people, according to the FBI Uniform Crime Report. There were 19,252 homicides in the United States in 2023. Those are real numbers, compiled from whatever data actually made it to the federal level. But they represent a floor constructed from partial agency participation and voluntary reporting, not a complete accounting of what happened in American streets last year.

Geography Is Not Random — And That Is the Point

Here is what makes this particularly brutal: crime does not distribute itself evenly. Research reviewed by the National Research Council found that roughly 50 percent of all crime concentrates in approximately 5 percent of street blocks. Not zip codes. Not neighborhoods. Street blocks. And those locations stay hot year after year, across decades. The same tiny fraction of physical geography generates the same outsized share of incidents. If you live on or near one of those blocks, the statistical noise around you is not noise. It is signal. Persistent, documented, predictable signal.

The system was not designed to give you that information. The annual crime statistics release, the neighborhood safety scores, the crime heat maps embedded in real estate apps — they aggregate, they average, they smooth. They were designed to satisfy policy discussions and budget hearings, not to tell you what is happening on your specific block. There is nothing conspiratorial about this. It is just a design problem. The tools that exist were built for different purposes, and you are using them to make decisions they were never engineered to support.

What You Think You Know

We trust proximity. We trust quiet. We trust the absence of visible disorder as evidence of actual safety. None of those are data. Quiet is not a statistic. The absence of crime tape is not a clearance rate. The neighborhood app where someone posted about their missing tabby is not a crime database. It is a social network with positive-feedback dynamics that reward the mundane and suppress the threatening.

The crime map you pulled up when you were house hunting? It shows reported crimes, from agencies that submitted data, within the current reporting year. That is three successive filters applied before you even load the page. You were looking at a projection of a projection.

This Is the Problem SafeScore Exists to Solve

Not to reassure you. Not to generate a green checkmark next to your zip code. The problem is that nobody has ever taken all of the available data — every public crime dataset, every agency submission, every geographic pattern in the historical record — and assembled it into a single, address-level picture that accounts for what the official numbers cannot capture on their own. The raw data exists. It has always existed. It just sits in disconnected federal databases, updated on different schedules, formatted for different purposes, accessible to anyone with a FOIA request and six months to spare.

We built a system that does the assembly for you. The score is not a guarantee. It is not a prediction. It is the most accurate possible reading of what the available record actually shows about one address — without the smoothing, without the aggregation, without the filter bubble of the neighborhood app. You get the floor of what is known. You decide what to do with it.

The comfortable picture of your neighborhood was never real. The map was always incomplete. Now you can see where the gaps are.

Related: What the Government Already Knows About Your Address →