How to use Airmid Éireann, Ireland's open pharma data: a guide for everyone

What Airmid Éireann is, in plain words

Airmid Éireann is a free collection of public data about medicines, clinical trials, health services and the pharmaceutical industry, built around Ireland. It takes records that official bodies already publish, such as the CSO, the HPRA, the European Medicines Agency and the US FDA, and puts them in one place, in the same simple format, with the source and licence of every file written down.

Without it, anyone who wants this information has to find dozens of websites, download spreadsheets, XML files and tables in different layouts, clean them, and then work out whether they are allowed to reuse them. Airmid Éireann does that work once, every week, and shares the result with everyone.

It is run by Priya Life Science and named after Airmid, the healer of Irish mythology who sorted the healing herbs of Ireland. The Airmid Éireann project page tells that story; this guide is about using the data.

How Ireland's open pharma data is organised

Each dataset sits in its own folder with the same three files: a CSV file you can open in Excel, a JSON file for programs, and a datapackage.json that names the publisher, the licence, the date the data was collected and every column. One index file, catalog.json, lists every dataset, its record count and the address of each file, with a SHA-256 checksum so you can check a download is complete.

The address of a file never changes between releases. The CSV of EU medicine shortages, for example, is always at the same address:

https://priyalifescience.com/opendata/airmid/v1/ema_shortages/ema_shortages.csv

Swap the dataset key (here ema_shortages) for any key in the table below and the pattern holds for all 46 datasets. Each release also includes a data dictionary that explains every column, a changelog, and ready-made citation files in BibTeX and RIS.

What is in the data: all 46 datasets

The table lists every dataset in the current release, grouped by topic, with its record count, its publisher and its licence. The licence matters: each dataset keeps the licence of the body that published it, so they are not all the same.

Two datasets are collected but not yet released: Department of Enterprise employment permit statistics and the Annual Employment Survey. They will be added once the Department confirms the licence in writing.

Who can use Airmid Éireann, and what for

The short answer is anyone who needs facts about medicines and health in Ireland and does not have the time or budget to collect them. The sections below take each group in turn, with the datasets to start with and a real question the data answers. None of it is medical advice: for decisions about your own health, talk to your doctor or pharmacist.

Patients, families and patient groups

Patients use the data to find trials, to check shortages and to understand waiting times. The trials data lists every study on ClinicalTrials.gov with a site in Ireland, with its status and conditions; in the release of 20 October 2026, 302 of them were recruiting. The EMA shortages file says which medicines have an ongoing EU shortage and whether alternatives exist.

• Start with: the search page, which finds records by meaning: type "recruiting breast cancer trial Dublin" or "insulin shortage" in plain words.
• Then: open the trial on ClinicalTrials.gov from the link in the record, and take it to your consultant. Only your care team can say whether a trial suits you.
• Patient groups: download ntpf_waiting_lists to track waiting times in your specialty month by month, and quote the NTPF as the source.

Journalists and fact-checkers

Journalists use Airmid Éireann to check a claim in minutes instead of days. Every record carries its source, so a figure can be traced back to the publisher before it goes to print.

• Shortages: ema_shortages and fda_shortages for the EU and US picture; read with our medicine shortages tracker.
• Politics: oireachtas_medicine_questions holds 10,323 parliamentary questions on medicines since 2020, with a link to the official record of each. In the current release, 54 questions asked in 2026 so far mention a shortage.
• Public money: etenders_health_awards shows health tenders and awards by the HSE and hospitals, with the winning companies.
• Industry: cso_pharma_exports shows where Irish pharma exports go. In the twelve months to August 2026, exports to the USA were worth €39.3 billion, the largest of any country, by the CSO's figures.
• Environment: eea_industrial_emissions_ireland lists pollutant releases and waste transfers reported by Irish chemical and pharmaceutical plants since 2007.

When you publish, credit the original publisher (for example "Source: CSO, via Airmid Éireann") and link the dataset.

Students, academics and researchers

Researchers get clean, documented data with a permanent DOI, which saves the weeks usually spent collecting and cleaning, and makes the method easy to check. An MSc thesis on Irish clinical trial activity, a PhD chapter on medicine shortages, or a paper on research funding can all start from a file that a reviewer can download and repeat.

• Research output and funding: openalex_ireland_pharma (Irish pharmacology and pharmaceutics papers), research_ireland_grants, cordis_health and nih_ireland.
• Health services research: cso_hospital_discharges (discharges and bed days by diagnosis and procedure), cso_health_accounts, ntpf_waiting_lists.
• Cross-border comparisons: ni_bso_gp_prescribing and nhs_prescription_costs use the same BNF substance codes, so the same medicine can be compared across Northern Ireland and England (see example 5 below).
• Cite the exact version you used (see "How to cite" below), so others can repeat your work with the same numbers.

Pharma, biotech and medtech companies

Companies use the data for market research and competitor tracking without paying for a commercial database. Business development, regulatory affairs, market access and supply chain teams each have a starting point.

• Regulatory and market access: ema_medicines, ema_prime, ema_paediatric_plans, eu_orphan_exclusivity (when orphan market exclusivity ends), and hpra_medicines for the Irish register.
• Patents and competition in the US: fda_orange_book (patents and exclusivities) and fda_purple_book (biologics and biosimilars).
• Clinical development: ctgov_ireland and ctgov_europe show who runs trials, where and in which conditions.
• Quality and compliance: fda_warning_letters, fda_device_recalls, mhra_safety_alerts and ema_referrals.
• Selling to the Irish health service: etenders_health_awards shows what was tendered, the value and the number of bids.

Quality, regulatory and pharmacovigilance professionals

QA, RA and safety staff use Airmid Éireann to keep watch on what regulators publish. The FDA warning letters file lists the subject, office and dates of each letter; device recalls give the classification and root cause; EMA PSUSA outcomes and referrals show EU safety decisions by active substance. Pair it with our guide to clinical trials in Ireland for the trial side, and with the methodology page to see how each register is built.

Pharmacists and GPs

Pharmacists check ema_shortages for the EU status of a medicine in short supply and hpra_medicines for whether a product is authorised and marketed in Ireland. Prescribing trends in Northern Ireland (ni_bso_gp_prescribing) show monthly items and cost by medicine for the whole region, which is useful context for practice teams on the island. Patient-level data is never included.

Policymakers, public bodies and the HSE

Officials, Oireachtas researchers and TDs use the data as a shared evidence base. Health spending by function and financing scheme (cso_health_accounts), the Department of Health's Key Trends tables (GMS prescriptions, medical cards, the hospital workforce), hospital discharges, waiting lists and parliamentary questions sit side by side, so a briefing can draw on several sources at once. International comparisons come from oecd_pharma_spending, eurostat_pharma_industry and ecdc_antimicrobial_consumption.

Investors and analysts

Analysts follow approvals, pipelines and trade. EMA approvals and PRIME designations show what is coming to the EU market; FDA approvals cover 25,405 US applications; Eurostat and CSO trade data show Ireland's pharma exports by month and partner country. Read our drug spending tracker for the US Medicaid and NHS context.

Environmental groups and local communities

People living near a plant can see what it reports. eea_industrial_emissions_ireland lists releases to air and water and waste transfers by facility and year, and epa_pharma_licences lists the EPA licences of pharmaceutical sites, with a link between the two through the EPA licence number. The Ireland pharma map shows where the plants are.

Developers, data scientists and people building AI tools

Developers get a stable, versioned API with no key: catalog.json for discovery and CSV or JSON for the data, served with open cross-origin access so a web page can read it directly. The code that builds the data is open on GitHub, so you can run it yourself, add a collector, or check exactly how a figure was produced. The sections below show how.

Teachers and trainers

The files are small enough to open in Excel and real enough to teach with: a statistics class can chart Irish pharma exports by month, a public health module can compare waiting lists by specialty, and a data course can practise joining tables on shared codes.

How to use the data without any coding

Search by meaning

1. Open the Airmid Éireann search page.
2. Type a question in plain words, for example "shortage of antibiotics for children" or "Phase 3 trial Cork".
3. The first search downloads a small language model into your browser, which takes a little time; later searches are fast. Your search stays on your device.
4. Each result links to the source record, so you can check it with the publisher.

Download a file and open it in Excel

1. Go to priyalifescience.com/opendata/airmid/v1 and choose a dataset.
2. Click the CSV file to download it, then open it in Excel, LibreOffice or Numbers.
3. If accents look wrong (for example in "Éireann"), use Data, then From Text/CSV, and choose UTF-8 as the file origin.
4. To keep the sheet up to date, use Data, then Get Data, then From Web, and paste the CSV address. Refresh the query each week to pick up the new release.

Google Sheets

For the smaller datasets, one formula loads the latest file into a sheet, and it refreshes on its own:

=IMPORTDATA("https://priyalifescience.com/opendata/airmid/v1/ema_shortages/ema_shortages.csv")

Large files (over a few megabytes) are better downloaded and imported, because spreadsheet tools limit how much they import from the web.

Power BI, Tableau and other dashboard tools

Choose "Get data from web" (or the tool's equivalent), paste the CSV address of the dataset, and schedule a weekly refresh to match the release cycle.

How to use the data in code

Every example below was run against the 20 October 2026 release before this guide was published. Copy them as they are; they read straight from the website.

Example 1: list every dataset (Python, nothing to install)

import json, urllib.request

BASE = "https://priyalifescience.com/opendata/airmid/v1/"
cat = json.load(urllib.request.urlopen(BASE + "catalog.json"))
print(cat["release"], len(cat["datasets"]), "datasets")
for d in cat["datasets"]:
    print(d["key"], d["records"], d["publisher"][:40])

Example 2: ongoing EU medicine shortages (Python with pandas)

Install pandas once with pip install pandas, then:

import pandas as pd

BASE = "https://priyalifescience.com/opendata/airmid/v1/"
shortages = pd.read_csv(BASE + "ema_shortages/ema_shortages.csv")
ongoing = shortages[shortages["status"] == "Shortage ongoing"]
print(len(ongoing), "ongoing shortages")
print(ongoing[["medicine", "inn", "therapeutic_area", "first_published"]])

In the 20 October 2026 release this prints 27 ongoing shortages. The other status values are "Shortage resolved" and "Medicine discontinued".

Example 3: where Irish pharma exports go

import pandas as pd

BASE = "https://priyalifescience.com/opendata/airmid/v1/"
exports = pd.read_csv(BASE + "cso_pharma_exports/cso_pharma_exports.csv")
last12 = exports[exports["month"] > "2025-08"]          # the twelve months to August 2026
top = last12.groupby("country")["exports_eur_thousand"].sum().sort_values(ascending=False)
print((top.head(10) / 1_000_000).round(2))                  # € billion

Values are in thousands of euro, as the CSO publishes them, so dividing by a million gives billions.

Example 4: trials recruiting in Ireland, by condition

import pandas as pd

BASE = "https://priyalifescience.com/opendata/airmid/v1/"
trials = pd.read_csv(BASE + "ctgov_ireland/ctgov_ireland.csv")
recruiting = trials[trials["overall_status"] == "RECRUITING"]
conditions = recruiting["conditions"].str.split("; ").explode()
print(len(recruiting), "recruiting")
print(conditions.value_counts().head(10))

Example 5: the same medicine in Northern Ireland and England

Both prescribing datasets use the nine-character BNF substance code, so they join directly. This compares atorvastatin, code 0212000B0, in one month:

import pandas as pd

BASE = "https://priyalifescience.com/opendata/airmid/v1/"
codes = {"bnf_chemical_substance_code": str}
ni = pd.read_csv(BASE + "ni_bso_gp_prescribing/ni_bso_gp_prescribing.csv", dtype=codes, low_memory=False)
en = pd.read_csv(BASE + "nhs_prescription_costs/nhs_prescription_costs.csv", dtype=codes)
month, code = "2026-07", "0212000B0"
ni_items = ni[(ni["year_month"] == month) & (ni["bnf_chemical_substance_code"] == code)]["items"].sum()
en_items = en[(en["year_month"] == month) & (en["bnf_chemical_substance_code"] == code)]["items"].sum()
print(month, "Northern Ireland", ni_items, "England", en_items)

For 2026-07 this gives 163,590 items in Northern Ireland and 7,091,022 in England. To compare fairly, divide each by its population from the official statistics office.

Example 6: parliamentary questions that mention shortages

import pandas as pd

BASE = "https://priyalifescience.com/opendata/airmid/v1/"
pq = pd.read_csv(BASE + "oireachtas_medicine_questions/oireachtas_medicine_questions.csv")
hits = pq[pq["question_text"].str.contains("shortage", case=False, na=False)]
print(hits.groupby("year").size())
print(hits[["date", "topic", "web_url"]].tail(10))

Example 7: download every file and check it

import hashlib, json, pathlib, urllib.request

BASE = "https://priyalifescience.com/opendata/airmid/v1/"
cat = json.load(urllib.request.urlopen(BASE + "catalog.json"))
folder = pathlib.Path("airmid-data")
for d in cat["datasets"]:
    for f in d["files"]:
        if f["format"] != "csv":
            continue
        data = urllib.request.urlopen(f["url"]).read()
        assert hashlib.sha256(data).hexdigest() == f["sha256"], "incomplete download: " + f["url"]
        path = folder / d["key"] / f["url"].rsplit("/", 1)[-1]
        path.parent.mkdir(parents=True, exist_ok=True)
        path.write_bytes(data)
        print("saved", path)

JavaScript, in a web page or Node.js

The files are served with open cross-origin access, so a web page on any site can read them directly:

const BASE = "https://priyalifescience.com/opendata/airmid/v1/";
const catalog = await (await fetch(BASE + "catalog.json")).json();
console.log(catalog.release, catalog.datasets.length, "datasets");
for (const d of catalog.datasets) console.log(d.key, d.records);

The command line

curl -s https://priyalifescience.com/opendata/airmid/v1/catalog.json \
  | jq -r '.datasets[] | [.key, .records] | @tsv'
curl -O https://priyalifescience.com/opendata/airmid/v1/ntpf_waiting_lists/ntpf_waiting_lists.csv

R

base <- "https://priyalifescience.com/opendata/airmid/v1/"
waits <- read.csv(paste0(base, "ntpf_waiting_lists/ntpf_waiting_lists.csv"), encoding = "UTF-8")
head(waits)

How to use Airmid Éireann with AI tools such as Claude and ChatGPT

AI assistants work well with this data because every file is a clean table with documented columns. There are three ways to connect them, from the simplest to the most capable.

1. Upload a file to the chat

1. Download the CSV you need (start with a small one, such as ema_shortages).
2. In Claude or ChatGPT, attach the file to a new chat.
3. Ask a plain question, and ask it to show its working, for example:

This CSV is the EMA medicine shortages dataset from Airmid Éireann
(source: European Medicines Agency). List the ongoing shortages by
therapeutic area, show the count for each, and tell me which column
you used for the status. Do not guess any values that are not in the file.

2. Give the assistant the catalogue address

Assistants that can read web pages can start from the index:

Read https://priyalifescience.com/opendata/airmid/v1/catalog.json.
Which datasets cover medicine shortages, and what licence does each have?
Then open the data dictionary for the one about the EU and explain its columns.

3. Connect Claude Desktop to a folder of Airmid data

Claude Desktop can read files on your computer through a Model Context Protocol (MCP) server. Download the datasets you want into one folder (example 7 above does it in one go), then:

1. Install Node.js if you do not have it.
2. In Claude Desktop, open Settings from the Claude menu, choose Developer, then Edit Config. This opens claude_desktop_config.json.
3. Add the filesystem server with the path to your data folder (Windows example; on a Mac use a path such as /Users/you/airmid-data):

   {
     "mcpServers": {
       "airmid-data": {
         "command": "npx",
         "args": [
           "-y",
           "@modelcontextprotocol/server-filesystem",
           "C:\\Users\\you\\airmid-data"
         ]
       }
     }
   }

4. Save the file and restart Claude Desktop. Claude can now open the files in that folder when you ask, for example "compare the waiting list totals in ntpf_waiting_lists.csv for the last two snapshots".

Give the server access to that data folder only. The official MCP quick start has screenshots of each step.

Rules for AI answers

• Check any number the assistant gives against the file before you use it.
• Cite the dataset and its original publisher, not the AI.
• Do not use AI answers from this data as medical advice.

How to run the collectors yourself from GitHub

The code on GitHub collects every dataset straight from its publisher, checks it and builds the release. It needs Python 3.10 or later and nothing else: no extra packages.

git clone https://github.com/priyalifescience/airmed
cd airmed
python3 -m airmed list                                # every source, with its status
python3 -m airmed run ema_shortages --out out          # collect one dataset
python3 -m airmed run all --out out                    # collect everything (large downloads)
python3 -m airmed release --src out --dist dist/v1     # build a release folder

The release step refuses data that is broken, much smaller than last time, or contains anything that looks like contact details. To keep your own copy fresh on a server, schedule the same commands weekly, for example with cron:

30 4 * * 1  cd ~/airmed && python3 -m airmed run all --out out && python3 -m airmed release --src out --dist dist/v1

Every source must be listed in sources/registry.json with its licence before a collector will run. To suggest a new source or report an error, open an issue or a discussion on GitHub.

How to cite Airmid Éireann

Cite the version you used, and credit the original publisher of each dataset. For the whole project:

Prasannan, S. (2026). Airmid Éireann: Ireland's Open Pharma Data Project (Version 1.3.0) [Data set and software]. Priya Life Science. https://doi.org/10.5281/zenodo.23499798

• All versions: DOI 10.5281/zenodo.23336432 always points to the latest release.
• One exact version: version 1.3.0 is 10.5281/zenodo.23499798. Use this in a paper or thesis so readers get the same numbers.
• Reference managers: import citation.bib or citation.ris.
• The original source: each datapackage.json holds the exact credit line, for example "Source: Central Statistics Office" for CSO data.

Licences, privacy and what is never included

Each dataset keeps the licence of its publisher, and each file states it. Most are CC BY 4.0, the UK or Canadian Open Government Licence, CC0 or public domain. A few carry their own terms, which you must follow when you reuse them:

• NIH RePORTER data is under the Open Database Licence (ODbL 1.0): if you share an adapted version, share it under the same licence.
• ClinicalTrials.gov data follows the National Library of Medicine's terms: credit ClinicalTrials.gov, give the date the data was processed, and say what you changed.
• OECD, NTPF and Eurostat data follow their own reuse terms, set out in each dataset's notes; the NTPF, for example, does not allow its data to be used mainly to advertise a product.
• Oireachtas material is reused under the Oireachtas open data licence, with its disclaimer.

No personal names or contact details are ever published: not of patients, researchers, members of the Oireachtas, sole traders or company staff. Data that we were licensed to use only on our own pages, such as NICE appraisals, is not shared. Errors we find are fixed and logged on the corrections page.

What changes with each release

This guide is updated with every Airmid Éireann release, so the counts, examples and table above always describe the current data.

• Version 1.3.0, 20 October 2026: 46 datasets. New: CSO health accounts, hospital discharges, Department of Health Key Trends tables, EEA industrial emissions of Irish plants, eTenders health awards, Oireachtas questions on medicines, Research Ireland grants, and Northern Ireland GP prescribing. Every licence was checked against the publisher's own terms.
• Version 1.2, 18 and 19 October 2026: 38 datasets, adding patents, biologics, Canada, EMA programmes, spending and waiting lists.
• Version 1.1, 16 October 2026: 19 datasets, adding Europe-wide trials, trade, industry, research and safety data.
• Version 1.0, 13 October 2026: first public release.

The full list of changes is in the release changelog, and our other registers are listed on the data trackers page.