Skip to main content
lab-notes.aiPublic-service AI playbooks
Opportunity 08/17JusticeJustice and community safety

Justice Research and Analysis

Explore which research questions published court figures can support when the available data is limited to quarterly offence-group counts.

Inspect starter dataSet up the build partner
Published sources
2
Starter data
20 synthetic records
Domain build partner
build-justice-research
Last reviewed

From the draft strategy

Opportunity

The draft strategy names justice data analysis and research as a potential public-service application, processing large volumes of material more quickly, without saying what access it assumes, what the research would be for, or which outputs would be acceptable.

Where the draft says it

Table 2 — potential public-service applications

Read the draft strategy source (opens in a new tab)

Source register

Research already done

Real published source

We reviewed these published sources so you do not have to start from zero. Each source shows what it covers, how it can be accessed, and why it matters.

2 sources

  1. Department of Justice

    Open

    Court Prosecutions, Convictions and Out of Court Disposals Statistics for Northern Ireland

    What it covers
    The yearly bulletin on prosecutions and convictions in Northern Ireland, broken down by court type, offence category, and disposal, with out of court disposals reported the same way.
    Why it matters here
    It is the published shape this playbook's synthetic records imitate: offences grouped into categories and counted, with nothing that identifies a case or a person.

    Open the source (opens in a new tab)

  2. Department of Justice

    Open

    Statistics and Research

    What it covers
    The department's collection of justice statistics and research: court and tribunal statistics, prosecutions and case processing times, prison service statistics, reoffending bulletins, the Safe Community Survey, youth justice workload, and research with victims and witnesses.
    Why it matters here
    It shows how much of the picture sits in separate publications on different cycles, which is most of the actual work in a question like this.

    Open the source (opens in a new tab)

Starter data

Starter dataset

Synthetic working data

Use 20 synthetic quarterly rows across five offence groups to frame research questions without case or person-level data.

AI-assisted research helped identify and interpret the published sources. We then created a small, non-sensitive synthetic dataset shaped by the information those sources expose. It is for exploration—not evidence, training, or operational decisions.

Records
20
Fields
5

Field preview

table view
  • id

    SYN-JR-01 · SYN-JR-02 · SYN-JR-03

  • offenceGroup

    Violence against the person · Theft · Criminal damage

  • quarter

    2025-Q1 · 2025-Q2 · 2025-Q3

  • disposalsBand

    500-1000 · 1000-2500 · 100-500

  • medianDaysToDisposal

    480 · 495 · 510

Inspect all 20 recordsView repository file (opens in a new tab)Download JSON (opens in a new tab)

How it was prepared

AI authored fictional banded disposal counts and median days to disposal for each offence group.

Limitations

  • The counts and the day figures are invented, so no row here describes what happens in any Northern Ireland court.
  • Five offence groups over one year cannot show the linkage error, missing records, and changes of process that a real analysis spends most of its time on.
  • A count of disposals says nothing about what happened to anyone, and a pattern found in a table like this cannot be turned into a statement about a person.

Clone and build

Domain build partner

Build with a domain-aware coding agent

Clone the repository, then ask your coding agent to load the skill that ships with it. The skill brings the opportunity, sources, starter data, known unknowns and constraints into the conversation before anything is proposed. It follows the open Agent Skills standard, so it works in Claude Code, Codex and any other agent that reads skills.

01 / Clone

git clone https://github.com/Hypership-Software/lab-notes.ai.git

02 / Invoke

Claude Code
/build-justice-research
Codex
$build-justice-research

Conditions, not a checklist

Before you build

These are not footnotes. They are conditions that any responsible prototype would need to address.

  • Justice records are sensitive and consequential

    Justice records can be sensitive, stigmatising, incomplete, and consequential all at once, which is why nothing person-level appears here.

  • Recorded patterns reflect institutional action

    Patterns in these records reflect who gets reported, stopped, and prosecuted as much as anything about behaviour, and anything trained on them learns that too.

  • Models can invent legal rules

    A model asked about law can invent a rule that sounds right, or leave out the exception that decides the case.

  • Exploratory patterns are not causes

    An exploratory pattern is not a cause, and it must never become a reason to do something to an individual.