🌆 Unique property reference numbers 🔢

💗 a very British love affair 💗

Addresses: a Data Quality Triple-threat

For “falsehoods programmers believe about addresses”, see Michael Tandy (2013); Paul Plowman (2018)

Addresses: a Data Quality Opportunity?

If Digital ID doesn’t play out it would be a shame for public services, but fret not…
- We can still just use name and address like always, but we do need to work hard to to raise the validity and consistency qualities of addresses stored in public service records.

💗 British locations‽
💗 British Standard 7666!

The kitemark is **not** used to certify the standard

Use British Standard 7666 to ensure that property, street, and land records are documented consistently (BSI Group 2021)

💌 Part 1/3 The beauty of UPRNs

UPRNs are <12 digit, trillion-sized sequential IDs assigned in blocks by Geoplace to councils who apply them. Their length makes them hard for humans to recall. Mnemonics may fix this, but plurals and homophones are a problem (joeldn 2025; BBC news 2021).

---
config:
    packet:
        showBits: false
        bitsPerRow: 40
        rowHeight: 70
---
packet-beta
0-39: "Geovation, Fourth Floor North, 65 Goswell Road, London, EC1V 7EN"
40-79: "10094608166"
80-89: "010"
90-99: "094"
100-109: "608"
110-119: "166"
120-129: "Pineapple"
130-139: "Latte"
140-149: "Corporal"
150-159: "Anchovy"
160-199: "1000 word dictionary"

💌 Going postal

UPRNs are published in AddressBase Premium, one of the nicer datasets you could hope to work with.

AddressBase Premium holds the address records from Royal Mail’s Postcode Address File, all address records from the Local Authority Land and Property Gazetteer, and objects without postal addresses (Ordnance Survey 2024).

💌 AddressBase attributes

💌 ~42 million UPRNs

💌 UPRNs give people an identity

Chris Hildrey (2020), Geomob Podcast (2020)

💌 UPRNs save lives

Skylar at Geomob (2023), Skylar at Fosdem (2025)

💌 UPRNs save lives

Shopkeepers sometimes create accommodation over a shop with a shuttered entrance that only unlocks from the outside!
Colleagues told me about a case where a tenant, now living in one such place, applied for a parking permit… at this point the parking team noted that the living quarters had not been registered as an address, inspected the property, and raised the alarm.
When serving residents, ensuring their address is genuine could save their life.

💌 UPRNs save lives

During lockdowns earlier this decade, the OS Open UPRN dataset, which contains only UPRNs + geospatial coordinates (not the commercial postal element), was shared with supermarkets for food parcel deliveries (Geoplace LLP 2020b).

OS Open UPRN will enable you to share and link data related to UPRNs, which you can spatially analyse and visualise using the accurate location (Ordnance Survey 2025).

💌 UPRN matching supports democracy

People who pay for democracy but haven’t joined yet.

SELECT * FROM 
  Taxpayers
LEFT JOIN 
  Voters
ON 
  Taxpayers.UPRN=Voters.UPRN
WHERE 
  Voters.UPRN IS NULL

💌 UPRN matching unlocks pupil premium

Lewisham’s Data Scientists unlocked service funding by linking school and welfare records (Policy in Practice 2025; Schoolsweek 2023)

💌 UPRN matching unlocks pension credits

To access welfare, pension status and pension credits (Policy in Practice 2023) are linked via UPRN
- Eligible households are supported to enrol by anti-poverty teams

💌 UPRN matching enables research

Three datasets linked via UPRN

School attendance
GP utilisation
National Child Measurement Programme (UK Government 2013...)

👹 Part 2/3 The beasts of BS7666 👺

UPRN (also USRN) was mandated in 2020 as a piece of critical UK infrastructure (see Geoplace LLP 2020a)

Yet [systems <~> integration] seems at an impasse…

🐉 Licensing Labyrinths (Nigel Shadbolt 2013; James O’Malley 2022-2025)

🐲 Software Silos (Mark Thompson & Jerry Fishenden 2013)

🧌 Procurement Pitfalls (Geoplace LLP 2025)

🛟 Part 3/3 Data matching to the rescue!

Systems integration remains valuable, but the practice of data matching has emerged as a data-driven form of DIY integration.

🎖️ https://docs.os.uk/os-apis/accessing-os-apis/os-places-api (single addresses only)

🥉 https://github.com/ONSdigital/aims-diy

🥈 https://www.robinlinacre.com/address_matching/

🥇 https://github.com/endeavourhealth-discovery/ASSIGN

🛟 ASSIGN: a uprn matching http api

AddreSS MatchInG to Unique Property Reference Numbers is made by Endeavour Health (David Stables 2013...)

ASSIGN is a highly accurate open-source address-matching algorithm with a high match rate and minimal biases when evaluated against a large sample of general practice-recorded patient addresses. ASSIGN has potential to be used in other address-based datasets (see Harper et al. 2019, 2021, 2024).

It also has the ability to de-identify datasets with a masking salt from OpenPseudonymiser (2011)

🛟 ASSIGN: Ensuring the currency of this highly dynamic dataset

📩 Each night, ASSIGN calls the OS API to check for a change-only update (COU) to AddressBase Premium

🌙 When there is a COU to process (frequently!), updates run overnight

🧇 Fresh data is served ~8 hours later

🛟 ASSIGN matching single addresses

---
config:
  theme: 'dark'
  messageFontSize: 20
---
sequenceDiagram
  %% actor You
  You ->> API : curl ... https://.../getinfo?adrec=Geovation,+Fourth+Floor+North,+65+Goswell+Road,+London,+EC1V+7EN
  API ->> You : {Address_format:good,Postcode_quality:good,Matched:true,<br/>BestMatch:{UPRN:🛖10093110833🛖,Qualifier:Property,LogicalStatus:1,<br/>Classification:PP,ClassTerm:Property Shell,Algorithm:130-bestfit,<br/>ABPAddress:{Number:65,Street:Goswell Road,Town:Islington,Postcode:EC1V 7EN,Organisaton:1},<br/>Match_pattern:{Postcode:equivalent,Street:equivalent,Number:equivalent,Building:candidate field dropped,Flat:equivalent}}}

🛟 ASSIGN matching multiple addresses

---
config:
  theme: 'dark'
---
sequenceDiagram
    %% actor You
    opt De-Identification Salt
      You -->> API : curl ... -F "file=test.encryptedSalt" https://.../fileupload2
      API -->> You : {"upload": { "status": "SALTOK"}}  
    end
    You ->> API : curl ... -F "file=addresses.txt" https://.../fileupload2
    API ->> You : {"upload": { "status": "OK"}}  
    You ->> API : curl ... "https://.../download3?filename=test-addresses.txt" --output "test-addresses.tsv"
    API ->> You : 💾 test-addresses.tsv

(Batches can be upto 100K records in size, matching is performed at ~38K records per minute)

🛟 Announcing the `assign-uprn` package

useful for: scientific notebooks; data pipelines
pre-requirements: a license to match (e.g. public sector geospatial); a permit from endeavour (url, user, password)
then it’s just three simple functions: address_search; upload; download

🛟 `address_search`

> address_search(
  'Geovation, Fourth Floor North, 65 Goswell Road, London, EC1V 7EN'
  ).json()
{'BestMatch': {
  'UPRN': '10093110833',
  'Qualifier': 'Property',
  'LogicalStatus': '1',
  'Classification': 'PP',
  'ClassTerm': 'Property Shell',
  'Algorithm': '130-bestfit',
  'ABPAddress': {'Number': '65',
   'Street': 'Goswell Road',
   'Town': 'Islington',
   'Postcode': 'EC1V 7EN',
   'Organisaton': '1'},
  'Match_pattern': {'Postcode': 'equivalent',
   'Street': 'equivalent',
   'Number': 'equivalent',
   'Building': 'candidate field dropped',
   'Flat': 'equivalent'}}
  'Address_format': 'good',
  'Postcode_quality': 'good',
  'Matched': True}

🛟 `upload`

test-addresses.txt:

1 ⭾ 10 Downing St, Westminster, London, SW1A2AA 
2 ⭾ Bridge Street, London, SW1A 2LW             
3 ⭾ 221b Baker St, Marylebone, London, NW1 6XE

comparable = {'upload': {'status': 'OK'}}
assert upload(
  infilepath="../data/external/test-addresses.txt"
  ).json() == comparable

comparable = {"upload": { "status": "SALTOK"}}
assert upload(
  infilepath="../data/external/test.EncryptedSalt"
  ).json() == comparable

🛟 `download`

infilepath = 'test-addresses.txt' # prior upload
outfilepath = '../data/processed/assign-uprn.tsv'
assert download(infilepath, outfilepath).status_code == 200

id	uprn	algorithm	classification	match_building	match_flat	match_number	match_postcode	match_street	abp_number	abp_postcode	abp_street	abp_town	qualifier	adr_candiddate	abp_building	latitude	longitude	point	x	y	ralf	classification_term	logical_status
1	100023336956	10-match1	RD04	equivalent	equivalent	equivalent	equivalent	equivalent	10	SW1A 2AA	Downing Street	City Of Westminster	Property	10 Downing St,Westminster,London,SW1A2AA		51.5035410	-.1276700	51.5035410	530047.00	179951.00	C30921C8404087803C3687301351FF41CCB4A5E8F3691070723293C8BD654CBB	Terraced	1
2	200002501505	550-match5a	PP	candidate field dropped	equivalent	equivalent	equivalent	equivalent		SW1A 2LW	Bridge Street	City Of Westminster	Property	Bridge Street,London,SW1A 2LW	Portcullis House	51.5013476	-.1243451	51.5013476	530284.00	179713.00	4D19E2EB66A2C12BD56B93D96CFBBE5B74525AEFC4C68329BE87B55C43EA4C36	Property Shell	1
3	100023071949	3200-match61A170	CR08	moved from Number	equivalent	moved to Building	equivalent	equivalent		NW1 6XE	Baker Street	London	Property	221b Baker St,Marylebone,London,NW1 6XE	221B	51.5237510	-.1585550	51.5237510	527847.00	182144.00	7727B90C7C3A744AF6FD8D5A4FEB6767B1EACBBC721B85EED6AE86EDD2B0BA9C	Shop / Showroom	1

🛟 Let’s match, clean, and de-identify

https://joeldn.srht.site/assign-uprn

THE END

BBC news. 2021. “App Used by Emergency Services Under Scrutiny — Bbc.co.uk.” https://www.bbc.co.uk/news/technology-56901363.

BSI Group. 2021. “BS 7666 - Spatial Datasets for Geographical Referencing — Landingpage.bsigroup.com.” https://doi.org/10.3403/BS7666.

Chris Hildrey. 2020. “ProxyAddress — Proxyaddress.org.” https://www.proxyaddress.org/.

David Stables. 2013... “Endeavour Health — Endeavourhealth.org.” https://endeavourhealth.org.

Geomob Podcast. 2020. “51. Chris Hildrey: ProxyAddress — thegeomob.com.” https://thegeomob.com/podcast/episode-51.

Geoplace LLP. 2020a. “Becoming Open and Royalty Free — Geoplace.co.uk.” https://www.geoplace.co.uk/power-of-place/becoming-open-and-royalty-free.

———. 2020b. “UPRNs – Supporting the Response to COVID-19 — Geoplace.co.uk.” https://www.geoplace.co.uk/blog/2020/uprns-supporting-the-response-to-covid-19.

———. 2025. “How to Generate Procurement Criteria — Geoplace.co.uk.” https://www.geoplace.co.uk/local-authority-resources/how-to/generate-procurement-criteria.

Harper, Gill, Nicola Firman, Marta Wilk, Milena Marszalek, Paul Simon, David Stables, Richard Fry, Kelvin Smith, and Carol Dezateux. 2024. “Determining Households from Patient Addresses and Unique Property Reference Numbers in General Practitioner Electronic Health Records.” International Journal of Population Data Science 9 (1): 2379.

Harper, Gill, David Stables, Kambiz Boomla, John Robson, Richard Fry, Zaheer Ahmed, and Carol Dezateux. 2019. “Allocating Unique Property Reference Numbers (UPRNs) to General Practitioner-Recorded Patient Addresses Using a Deterministic Address-Matching Algorithm: Evaluation of Representativeness and Bias in an Ethnically-Diverse Inner City Population.” International Journal of Population Data Science 4 (3).

Harper, Gill, David Stables, Paul Simon, Zaheer Ahmed, Kelvin Smith, John Robson, and Carol Dezateux. 2021. “Evaluation of the ASSIGN Open-Source Deterministic Address-Matching Algorithm for Allocating Unique Property Reference Numbers to General Practitioner-Recorded Patient Addresses.” International Journal of Population Data Science 6 (1): 1674.

James O’Malley. 2022-2025. “Odds and Ends of History | James O’Malley | Substack — Takes.jamesomalley.co.uk.” https://takes.jamesomalley.co.uk/t/paf.

joeldn. 2025. “Mnemonic-Place-Encoder — joeldn.srht.site.” https://joeldn.srht.site/site/post/mnemonic-place-encoder.html.

Mark Thompson & Jerry Fishenden. 2013. “The Great Deverticalisation – Part 1: The Street | Computer Weekly — Computerweekly.com.” https://www.computerweekly.com/opinion/The-great-deverticalisation-part-1-the-street.

Michael Tandy. 2013. “Falsehoods Programmers Believe about Addresses — Mjt.me.uk.” https://www.mjt.me.uk/posts/falsehoods-programmers-believe-about-addresses/.

Nigel Shadbolt. 2013. “A Cornerstone for Open Data: The Postcode Address File — Theodi.org.” https://theodi.org/news-and-events/blog/a-cornerstone-for-open-data-the-postcode-address-file/.

OpenPseudonymiser. 2011. “OpenPseudononymiser — Openpseudonymiser.org.” https://www.openpseudonymiser.org.

Ordnance Survey. 2024. “Which Address Product Should You Use? | More Than Maps — Docs.os.uk.” https://docs.os.uk/more-than-maps/os-mastermap-generation-data-demonstrators/addressing-and-location-demonstrators/which-address-product-should-you-use#addressbase-premium.

———. 2025. “OS Open UPRN | Data Products | OS — Ordnancesurvey.co.uk.” https://www.ordnancesurvey.co.uk/products/os-open-uprn.

Paul Plowman. 2018. “UK Address Oddities! &#X2013; The Unusually Named Blog — Paulplowman.com.” https://paulplowman.com/stuff/uk-address-oddities/.

Policy in Practice. 2023. “Low Income Family Tracker (LIFT) | Policy in Practice — Policyinpractice.co.uk.” https://policyinpractice.co.uk/low-income-family-tracker.

———. 2025. “Three London Boroughs Tackle Child Hunger by Auto Enrolment of 1,500 Students in Free School Meals | Policy in Practice — Policyinpractice.co.uk.” https://policyinpractice.co.uk/case-study/fsm/.

Schoolsweek. 2023. “Council’s £800 Exercise Results in £1.2m School Cash Boost — Schoolsweek.co.uk.” https://schoolsweek.co.uk/councils-800-exercise-results-in-1-2m-school-funding-boost/.

Skylar at Fosdem. 2025. “How to Save a Life.” https://archive.fosdem.org/2025/schedule/event/fosdem-2025-4914-how-to-save-a-life/.

Skylar at Geomob. 2023. “Past Geomob Speakers — Thegeomob.com.” https://thegeomob.com/past-speakers.

UK Government. 2013... “National Child Measurement Programme — Gov.uk.” https://www.gov.uk/government/collections/national-child-measurement-programme.

🌆 Unique property reference numbers 🔢

Addresses: a Data Quality Triple-threat

Addresses: a Data Quality Opportunity?

💗 British locations‽ 💗 British Standard 7666!

💌 Part 1/3 The beauty of UPRNs

💌 Going postal

💌 AddressBase attributes

💌 ~42 million UPRNs

💌 UPRNs give people an identity

💌 UPRNs save lives

💌 UPRNs save lives

💌 UPRNs save lives

💌 UPRN matching supports democracy

💌 UPRN matching unlocks pupil premium

💌 UPRN matching unlocks pension credits

💌 UPRN matching enables research

👹 Part 2/3 The beasts of BS7666 👺

🛟 Part 3/3 Data matching to the rescue!

🛟 ASSIGN: a uprn matching http api

🛟 ASSIGN: Ensuring the currency of this highly dynamic dataset

🛟 ASSIGN matching single addresses

🛟 ASSIGN matching multiple addresses

🛟 Announcing the assign-uprn package

🛟 address_search

🛟 upload

🛟 download

🛟 Let’s match, clean, and de-identify

THE END

💗 British locations‽
💗 British Standard 7666!

🛟 Announcing the `assign-uprn` package

🛟 `address_search`

🛟 `upload`

🛟 `download`