Bolster

A Python library for accessing Northern Ireland and UK government open data. Covers population, health, crime, economy, transport, housing, and more — each source normalised into a clean pandas DataFrame with a matching CLI command.

Full documentation: bolster.readthedocs.io

Installation

pip install bolster

Requires Python 3.11+.

What can you do with it?

A few examples using real data:

Who is claiming UC/JSA, and where?

from bolster.data_sources.nisra import claimant_count

df = claimant_count.get_latest_claimant_count("lgd")
# 33,930 claimants across NI as of May 2026
# Derry City & Strabane has the highest rate at 4.1%
print(
    df[df["date"] == df["date"].max()].sort_values("claimant_rate_total_pct", ascending=False)[
        ["geography", "claimants_total", "claimant_rate_total_pct"]
    ]
)

How many people are on the hypertension register at each GP practice?

from bolster.data_sources.health_ni import disease_prevalence

# NI-wide: 304,952 patients on the hypertension register in 2024/25
summary = disease_prevalence.get_latest_disease_prevalence(level="ni")

# GP-practice level: ~360 practices across ~17 financial years
gp = disease_prevalence.get_latest_gp_prevalence()
print(gp[gp["disease"] == "Hypertension"].groupby("financial_year")["registered_patients"].sum())

How does GP list size vary by Trust, and how has it changed?

from bolster.data_sources.health_ni import gms

df = gms.get_list_size(level="trust")
# Western Trust has the largest lists at 1,619.8 patients/GP in 2025/26,
# Belfast the smallest at 1,277.0 -- both up from five years earlier
latest = df[df["period"] == df["period"].max()].sort_values("list_size")
print(latest)

How has the security situation in NI changed since the Troubles?

from bolster.data_sources.psni import security_situation

df = security_situation.get_deaths()
# 470 deaths in 1972, the worst year on record; 0 in both 2023 and 2025 --
# a dramatic long-term decline, though the series hasn't reached zero every year
print(df[["year", "total"]].sort_values("total", ascending=False).head(3))
print(df[df["year"] >= 2020][["year", "total"]])

How is NI's economy tracking against pre-pandemic levels?

from bolster.data_sources.nisra import composite_index

df = composite_index.get_latest_nicei()
# NICEI at 105.9 in Q1 2026 (base 100 = 2016)
# Breakdown by sector: services, production, construction, agriculture
print(df.tail(4)[["year", "quarter", "nicei", "services", "production"]])

What are A&E waiting times doing across HSC Trusts?

from bolster.data_sources.health_ni import emergency_care_waiting_times

df = emergency_care_waiting_times.get_latest_data()
# Monthly attendance and % seen within 4 hours by Trust and department type
latest = df[df["date"] == df["date"].max()]
print(latest.groupby("trust")["pct_within_4hrs"].mean().sort_values())

What is the median full-time weekly wage in NI?

from bolster.data_sources.nisra import ashe

df = ashe.get_latest_ashe_timeseries("weekly")
# £713.10 median full-time weekly earnings in 2025
print(df[df["work_pattern"] == "Full-time"].tail(5)[["year", "median_weekly_earnings"]])

Data sources

Sources are organised by publisher. Each module follows the same pattern: get_latest_*() returns a tidy DataFrame; bolster <command> --help gives CLI access.

NISRA — NI Statistics and Research Agency

People and society: population, births, deaths, marriages, stillbirths, migration, population_projections, baby_names, registrar_general, deprivation, wellbeing, public_confidence, drug_related_deaths, armed_forces_veterans

Economy and labour: labour_market, claimant_count, ashe, quarterly_employment_survey, composite_index, index_of_production, index_of_services, construction_output, business_register, planning_statistics, housing_stock, housing_bulletin, tourism, work_quality, workless_households, neet

Education: teacher_workforce, school_leavers, highest_qualification

Department of Health NI (health_ni)

disease_prevalence (NI / LGD / HSCT / GP-practice level), cancer_waiting_times, diagnostic_waiting_times, elective_waiting_times, hospital_activity (beds, independent-sector activity, theatre usage), emergency_care_waiting_times, child_protection, hsc_workforce / hsc_recruitment, gms (GP practices/GPs/patients by trust, LGD or GP federation, funding, access equity)

PSNI — Police Service of Northern Ireland

crime_statistics (historical), road_traffic_collisions, stop_and_search, pace, police_ombudsman, motoring_offences, road_safety_partnership (safety camera detections), security_situation (deaths, incidents, paramilitary attacks, arrests, 1969-present), breath_tests (preliminary breath tests conducted, 2010-present)

Department of Justice (justice)

prosecutions_convictions, first_time_entrants, nicts_quarterly, mortgages, pbni_caseload, pps_statistical_bulletin

Other sources

CLI

Every data source has a matching CLI command:

bolster nisra deaths                        # latest weekly deaths
bolster nisra claimant-count --breakdown lgd
bolster nisra composite-index
bolster health-ni disease-prevalence --level gp
bolster psni stop-and-search
bolster translink departures "Europa Buscentre"
bolster water-quality BT1 5GS
bolster dva vehicle-tests
bolster --help                              # full command list

Utilities

Bolster also includes general-purpose helpers used internally:

• poolmap() — ThreadPoolExecutor with progress bar and error handling
• backoff() — exponential backoff retry decorator
• memoize() — instance method cache with hit/miss tracking
• get_recursively() / flatten_dict() — nested dict navigation
• CachedDownloader — disk-cached HTTP download with TTL
• session — shared requests.Session with retry/jitter logic

import bolster

results = bolster.poolmap(lambda x: x**2, range(1000), max_workers=4)


@bolster.backoff(Exception, tries=3, delay=1, backoff=2)
def unreliable(): ...

Development

git clone https://github.com/andrewbolster/bolster.git
cd bolster
uv sync --all-extras --dev
uv run pre-commit install
uv run pytest tests/ -q --no-cov        # quick run
uv run pytest tests/ --cov=src/bolster  # with coverage

See AGENTS.md for the data source development workflow and CONTRIBUTING.md for guidelines.

License

GNU General Public License v3