Building an address register for the
2021 Census
and beyond
Alistair Calder
Head of Address Register (ONS)
• Requirement – addresses in census – and why it’s easy
• Issues and our strategy – and why it’s hard
• The bigger picture – a national addressing service
• Our strategy & plan for 2021
CENSUS OPERATION 2021
CCS
Enforce
-ment
Build
register
ADDRESS
REGISTER
Post
Out
IAC
Hand
Deliver
HOUSE
HOLDS
COMMUNAL
ESTABLISHMENTS
100K ?
29M ?
ONLINE
Completion
Paper
??
Track
response
Follow
Up
Reminder
Letters
Emails
RESPONSE
DATABASE
Estimation
& outputs
Admin
data
OUTPUT
DATABASE
ESTIMATION
CENSUS OPERATION 2021
Post
Out
IAC
Hand
Deliver
ONLINE
Completion
Paper
??
CCS
Enforce
-ment
Track
response
Build
register
ADDRESS
REGISTER
Follow
Up
HOUSE
HOLDS
COMMUNAL
ESTABLISHMENTS
100K ?
29M ?
Estimation
& outputs
Admin
data
OUTPUT
DATABASE
Reminder
Letters
emails
RESPONSE
DATABASE
ESTIMATION
The requirement (to be agreed)
• A ‘complete’ household frame >99% of household
spaces
( addresses)
• Minimal over-coverage
duplicates / commercial property / demolished etc
< 2 or 3% ?
• A brilliant communal frame (integrated with
residential)
• Up to date, correctly located etc etc …. And more
Postcode
Address
File
PAF
Valuation
Office
National Land
& Property
Gazetteer
NLPG
Address
Layer 2
AL2
LLPGs
LLPGs
x 348
Council
Tax
TV
License
Utilities
Emerg.
Services
etc
Address Register
The olden days
Address Register – Census 2011
PAF
AL2
VOA
Residential address list
RULES CLERICAL
Communal address list
NLPG
Very good Just good enough
Why it’s going to be
easy this time ….
PAF
AL2
VOA
Residential address list
RULES
NLPG
Residential address list
PAF
AL2
VOA
RULES
NLPG
Communal address list
Additional CE sources
2011 Census AR AddressBase
Why it’s going to be
hard this time ….
The challenge ..... Why it’s hard this time
• We have an excellent starting point but addresses are
complicated and change a lot. There will be error & error
clusters itself in the areas we care about the most – Very
difficult to check quality
• Extracting the right ones is difficult. Small errors can be
significant – and cause trauma
• Communals are important and particularly challenging
• Addresses are complex so matching is really hard
• We plan to do MUCH more with the register than post-out –
huge opportunity but attribute thinking is new
The Emerging Strategy
what’s the plan?
Flat 1 Flat 2
Flat 3 Flat 4
Flat 5 Flat 6
Flat 7
7
The Emerging Strategy
what’s the plan?
42 5 5? B
10
The challenge ..... Why it’s hard this time
• We have an excellent starting point but addresses are
complicated and change a lot. There will be error & error
clusters itself in the areas we care about the most – Very
difficult to check quality
• Extracting the right ones is difficult. Small errors can be
significant – and cause trauma
• Communals are important and particularly challenging
• We plan to do MUCH more with the register than post-out –
huge opportunity but attribute thinking is new
• Addresses are complex so matching is really hard
The challenge ..... Why it’s hard this time
• We have an excellent starting point but addresses are
complicated and change a lot. There will be error & error
clusters itself in the areas we care about the most – Very
difficult to check quality
• Extracting the right ones is difficult. Small errors can be
significant – and cause trauma
• Communals are important and particularly challenging
• We plan to do MUCH more with the register than post-out –
huge opportunity but attribute thinking is new
• Addresses are complex so matching is really hard
Lists of
communals
Compared to
counts from
admin data
Linked to
Business
Index
Linked to
Address
Index
The challenge ..... Why it’s hard this time
• We have an excellent starting point but addresses are
complicated and change a lot. There will be error & error
clusters itself in the areas we care about the most – Very
difficult to check quality
• Extracting the right ones is difficult. Small errors can be
significant – and cause trauma
• Communals are important and particularly challenging
• We plan to do MUCH more with the register than post-out –
huge opportunity but attribute thinking is new
• Addresses are complex so matching is really hard
Where to check ??
Electricity - Multiple meters
Where to check ??
Electoral register - Multiple names
... also directly relevant topics such as internet access ... now updated opendata at postcode
Albert Road
Southsea
Common
A probabilistic address frame
Probability of
• Existence of address
• type - HH/B/CE
• HH Size / structure
• Change / churn
• Hard to countness / category
• (multivariate >> categorisation
• Eg possible holiday home, carehome, student
accommodation
Address
Register
HH
Structure
2011
Census
HH structure,
churn, names
Activity data
Energy, utilities,
broadband, health,
house sales
Admin data
HH structure, churn,
names, house
prices, phone
numbers
Other
Shape / pattern
recognition
Survey paradata
Geoplace
And other CE sources
CE
New definition / schema
Inform field planning / targeting
Intelligent stratification
Prioritise follow up (address level)
Inform estimation & modelling
B
Business Reg
Business structure,
type, churn
Conceptually – all subject to ethical and privacy discussion !
Potentially
The challenge ..... Why it’s hard this time
• We have an excellent starting point but addresses are
complicated and change a lot. There will be error & error
clusters itself in the areas we care about the most – Very
difficult to check quality
• Extracting the right ones is difficult. Small errors can be
significant – and cause trauma
• Communals are important and particularly challenging
• We plan to do MUCH more with the register than post-out –
huge opportunity but attribute thinking is new
• Addresses are complex so matching is really hard
•
Government Address Register
Service
Government
Address
Register
Service
Government Digital Service (GDS)
Vision for Registers
Address Index Alpha – Key Functions
ONS or
citizen
servicesingle
address UPRN
10 High St PO15 5RR 1234567891011
batch of
addresses
addresses
UPRNs
batch
match
Addressbase load
UPRNs
addresses
classifications
add attributes
UPRNs
addresses
classifications
attributes
extract
Feedback
to source
(improving quality)
api
api
api
Alpha ǀ Address Service Demo
Alpha ǀ Address Service Demo
Alpha – Address
Register build
2015 2016 2017 2018 2019 2020 2021 2022 2023April July October April July October
On-line Survey
transformation
Admin
Data
Admin Data –
Processing
Platform
Alpha
EDC – eQ Alpha EDC – eQ Beta
EDC – Response and Respondent Management Beta
Admin Data – Processing Platform Beta
EDC – Service enhancement
Admin/Survey
Integration
Discovery
Admin/Survey
Integration –
Alpha
Admin/Survey Integration – Beta
Alpha - Business
Register build
Beta - Business
Register build
Beta - Address
Register build
Registers
2019
Census
Rehearsal
Admin
Data for
Census
Census
Register Platform for ONS
Live services
Decision to
proceed to
beta Develop data migration and data loader for new
BIS data source
IDBR Service Migration
IDBR Migration
Roadmap
Business Statistics
Decision(s) to go
live
2021
Census
Life Events, Social Survey etc etc
Strategy on a page
Compare to other
sources
Field
check
Rehearsal
Test
Desk
check
Link sources
Admin &
commercial
Quality
Framework
For
communals
Best
sources
Graph
databases
Understand &
improve quality
of addressbase

Excellence in
matching
Government
Address
Register
Service

Fix
communals

Build
probablistic
register

citizen govt

Rules
• Address list ---> Probabalistic address register
• Address check ---> Desk work & checking
• Separate residential & CE lists ---> Integrated list
• Trauma ---> Proper field procedures for coping with error
• Unknown quality ---> High quality list cleaned by LAs
• ---> A National Address Register Service
2011 ---> 2021 – change

Building an address register for the 2021 Census and beyond

  • 1.
    Building an addressregister for the 2021 Census and beyond Alistair Calder Head of Address Register (ONS)
  • 2.
    • Requirement –addresses in census – and why it’s easy • Issues and our strategy – and why it’s hard • The bigger picture – a national addressing service • Our strategy & plan for 2021
  • 3.
    CENSUS OPERATION 2021 CCS Enforce -ment Build register ADDRESS REGISTER Post Out IAC Hand Deliver HOUSE HOLDS COMMUNAL ESTABLISHMENTS 100K? 29M ? ONLINE Completion Paper ?? Track response Follow Up Reminder Letters Emails RESPONSE DATABASE Estimation & outputs Admin data OUTPUT DATABASE ESTIMATION
  • 4.
  • 5.
    The requirement (tobe agreed) • A ‘complete’ household frame >99% of household spaces ( addresses) • Minimal over-coverage duplicates / commercial property / demolished etc < 2 or 3% ? • A brilliant communal frame (integrated with residential) • Up to date, correctly located etc etc …. And more
  • 6.
    Postcode Address File PAF Valuation Office National Land & Property Gazetteer NLPG Address Layer2 AL2 LLPGs LLPGs x 348 Council Tax TV License Utilities Emerg. Services etc Address Register The olden days
  • 7.
    Address Register –Census 2011 PAF AL2 VOA Residential address list RULES CLERICAL Communal address list NLPG Very good Just good enough
  • 8.
    Why it’s goingto be easy this time ….
  • 9.
    PAF AL2 VOA Residential address list RULES NLPG Residentialaddress list PAF AL2 VOA RULES NLPG Communal address list Additional CE sources 2011 Census AR AddressBase
  • 10.
    Why it’s goingto be hard this time ….
  • 11.
    The challenge .....Why it’s hard this time • We have an excellent starting point but addresses are complicated and change a lot. There will be error & error clusters itself in the areas we care about the most – Very difficult to check quality • Extracting the right ones is difficult. Small errors can be significant – and cause trauma • Communals are important and particularly challenging • Addresses are complex so matching is really hard • We plan to do MUCH more with the register than post-out – huge opportunity but attribute thinking is new
  • 14.
    The Emerging Strategy what’sthe plan? Flat 1 Flat 2 Flat 3 Flat 4 Flat 5 Flat 6 Flat 7
  • 15.
  • 16.
  • 19.
    42 5 5?B 10
  • 20.
    The challenge .....Why it’s hard this time • We have an excellent starting point but addresses are complicated and change a lot. There will be error & error clusters itself in the areas we care about the most – Very difficult to check quality • Extracting the right ones is difficult. Small errors can be significant – and cause trauma • Communals are important and particularly challenging • We plan to do MUCH more with the register than post-out – huge opportunity but attribute thinking is new • Addresses are complex so matching is really hard
  • 25.
    The challenge .....Why it’s hard this time • We have an excellent starting point but addresses are complicated and change a lot. There will be error & error clusters itself in the areas we care about the most – Very difficult to check quality • Extracting the right ones is difficult. Small errors can be significant – and cause trauma • Communals are important and particularly challenging • We plan to do MUCH more with the register than post-out – huge opportunity but attribute thinking is new • Addresses are complex so matching is really hard
  • 26.
    Lists of communals Compared to countsfrom admin data Linked to Business Index Linked to Address Index
  • 27.
    The challenge .....Why it’s hard this time • We have an excellent starting point but addresses are complicated and change a lot. There will be error & error clusters itself in the areas we care about the most – Very difficult to check quality • Extracting the right ones is difficult. Small errors can be significant – and cause trauma • Communals are important and particularly challenging • We plan to do MUCH more with the register than post-out – huge opportunity but attribute thinking is new • Addresses are complex so matching is really hard
  • 28.
    Where to check?? Electricity - Multiple meters
  • 29.
    Where to check?? Electoral register - Multiple names
  • 30.
    ... also directlyrelevant topics such as internet access ... now updated opendata at postcode Albert Road Southsea Common
  • 31.
    A probabilistic addressframe Probability of • Existence of address • type - HH/B/CE • HH Size / structure • Change / churn • Hard to countness / category • (multivariate >> categorisation • Eg possible holiday home, carehome, student accommodation Address Register HH Structure 2011 Census HH structure, churn, names Activity data Energy, utilities, broadband, health, house sales Admin data HH structure, churn, names, house prices, phone numbers Other Shape / pattern recognition Survey paradata Geoplace And other CE sources CE New definition / schema Inform field planning / targeting Intelligent stratification Prioritise follow up (address level) Inform estimation & modelling B Business Reg Business structure, type, churn Conceptually – all subject to ethical and privacy discussion ! Potentially
  • 32.
    The challenge .....Why it’s hard this time • We have an excellent starting point but addresses are complicated and change a lot. There will be error & error clusters itself in the areas we care about the most – Very difficult to check quality • Extracting the right ones is difficult. Small errors can be significant – and cause trauma • Communals are important and particularly challenging • We plan to do MUCH more with the register than post-out – huge opportunity but attribute thinking is new • Addresses are complex so matching is really hard •
  • 33.
  • 35.
    Government Digital Service(GDS) Vision for Registers
  • 36.
    Address Index Alpha– Key Functions ONS or citizen servicesingle address UPRN 10 High St PO15 5RR 1234567891011 batch of addresses addresses UPRNs batch match Addressbase load UPRNs addresses classifications add attributes UPRNs addresses classifications attributes extract Feedback to source (improving quality) api api api
  • 37.
    Alpha ǀ AddressService Demo
  • 38.
    Alpha ǀ AddressService Demo
  • 39.
    Alpha – Address Registerbuild 2015 2016 2017 2018 2019 2020 2021 2022 2023April July October April July October On-line Survey transformation Admin Data Admin Data – Processing Platform Alpha EDC – eQ Alpha EDC – eQ Beta EDC – Response and Respondent Management Beta Admin Data – Processing Platform Beta EDC – Service enhancement Admin/Survey Integration Discovery Admin/Survey Integration – Alpha Admin/Survey Integration – Beta Alpha - Business Register build Beta - Business Register build Beta - Address Register build Registers 2019 Census Rehearsal Admin Data for Census Census Register Platform for ONS Live services Decision to proceed to beta Develop data migration and data loader for new BIS data source IDBR Service Migration IDBR Migration Roadmap Business Statistics Decision(s) to go live 2021 Census Life Events, Social Survey etc etc
  • 40.
    Strategy on apage Compare to other sources Field check Rehearsal Test Desk check Link sources Admin & commercial Quality Framework For communals Best sources Graph databases Understand & improve quality of addressbase  Excellence in matching Government Address Register Service  Fix communals  Build probablistic register  citizen govt  Rules
  • 41.
    • Address list---> Probabalistic address register • Address check ---> Desk work & checking • Separate residential & CE lists ---> Integrated list • Trauma ---> Proper field procedures for coping with error • Unknown quality ---> High quality list cleaned by LAs • ---> A National Address Register Service 2011 ---> 2021 – change