Methods / limits / provenance

Numbers before claims.

The site is designed so a reader can tell what was counted, when it was observed, where it came from, and what the public record cannot prove.

Snapshot computedJul 9, 2026Mar 5-Jul 9, 2026

01 / Source

A public administrative record, mirrored locally.

The primary source is the Michigan Department of Corrections Offender Tracking Information System (OTIS). The local mirror retains source URLs and observation timestamps, then computes aggregate statistics for this publication.

1OTIS public pages

Administrative profiles and sentence fields.

2Local evidence mirror

142,592 current records plus dated snapshots.

3Aggregate builder

Counts, buckets, baselines, and coverage checks.

4Public data studio

No evidence files or private submissions in charts.

Michigan OTIS public search

02 / Scope and coverage

What the snapshot contains.

The Jul 9 aggregate contains 142,592 people and 520,942 sentence rows. A person can have multiple sentence rows, so row-level charts should never be read as unique-person counts.

FieldCoveredMissing / uncategorizedCoverage
Core record142,5920100.0%
Canonical status142,46812499.9%
Age band142,55240100.0%
Binary sex field142,56824100.0%
Maximum sentence518,6752,26799.6% of sentence rows

County jail populations, youth records, sealed or removed records, and court-file detail outside OTIS are out of scope. Source updates can lag real-world events.

03 / Definitions

Terms kept close to the chart.

People in the mirror
Distinct current OTIS records keyed by the local identity model. This is not the current prison population.
Marked Prisoner
Records whose current canonicalized status is Prisoner at the latest local observation.
Sentence row
One parsed prison or probation sentence entry. One person can contribute many rows.
Observed event
A change derived between local snapshots. Counts are events, not unique people or official statewide rates.
Population baseline
A resident-population comparison used to describe representation. It is not an explanation of cause.
Past ERD / PBJD
Marked Prisoner with an earliest release or parole-jurisdiction date on or before the as-of date. Eligibility is not guaranteed release.

04 / Privacy and participation

Public data and private contributions stay separate.

Aggregate charts

No user-submitted evidence or profile email enters the public data studio.

Private evidence

Uploaded files remain in protected object storage and are not publicly addressable.

Moderated publication

Permission to publish anonymously is optional and never bypasses review.

Server-side ownership

Profiles, submissions, and saved views are tied to the authenticated user on the server.

05 / Aggregate downloads

Take the reviewed rows with you.

These CSV files contain only the compact aggregate rows used in the visualizations.

A gap is a research question, not a license to guess.

Use the data studio to locate patterns, then use primary records and domain expertise before making a public claim.

Return to the data studio