Version: 1.0.7 | Published: 26 Nov 2025 | Updated: 15 days ago
Which? Priority Places for Insulation Index
Dataset
Summary
Short Description:
The Priority Places for Insulation Index (PPII) ranks geographical areas within
each nation of the UK across eight different indicators relating to insulation
needs.
Tier:
Tier 0
Access Type:
Open
Theme:
Sustainability
Documentation
Associated Media:
Description:
The Priority Places for Insulation Index (PPII) ranks geographical areas within each nation of the UK across eight different indicators relating to insulation needs. The index is produced by weighting and combining the eight indicators to construct an overall ranking of Priority Places for Insulation within each nation. It has been produced at three levels of geographical: local authorities, parliamentary constituencies and small local areas.
A web tool to display, map and graph the results for your local area can be found at https://insulation.which.co.uk/.
The index was developed by Which?. It was constructed using open data and web-scraped data from the publicly accessible Energy Performance Certificate (EPC) register for Northern Ireland as well as relevant sites where installers can be found or listed. The index is novel because it incorporates both aspects of housing stock and household circumstances which exacerbate the need for insulation.
How Published:
Data asset provided by the Healthy & Sustainable Places Data Service
(ES/Z504336/1), Created by Which? in partnership with CDRC
(ES/L011840/1;ES/L011891/1)
Coverage
Spatial
Spatial Units:
- Lower Super Output Areas (2011)
- Data Zones (2011)
- Super Output Areas (2011)
- Local Authority Districts (2022)
- Westminster Parliamentary Constituencies (2022)
Spatial Coverage:
United Kingdom
Temporal
Frequency:
STATIC
Date of Latest Release:
28 May 2025
Date of First Release:
29 November 2023
Temporal Aggregation:
Snapshot in time
Geographic Bounding Box
Lower Left Latitude:
-8.2
Lower Left Longitude:
49.9
Upper Right Latitude:
1.9
Upper Right Longitude:
61.0
Provenance
Source:
- Administrative data
- Web-scraped data
Purpose:
Data were originally collected for administrative purposes.
Collection Status:
Complete
Method of Collection:
To transform the data into rankings we used a methodology adapted from the
English Indices of Multiple Deprivation. For the health conditions indicator,
for which we have data for multiple diseases, steps 1-3 below were applied to
normalise the data for each disease rate before combining them to construct the
indicator. For other indicators, the procedure begins at step 4. 1. Each disease
rate is grouped by country and then ranked according from the highest to the
lowest values. 2. Each disease rate is transformed using the Rankit method to
normalise the data 3. Indicators are constructed by averaging over each disease
rate within each indicator 4. Each indicator is grouped by each country and then
ranked with ties from highest priority to lowest priority. 5. The indicator
ranks are transformed to an exponential distribution to minimise cancellation
effects. The exponential transformation applied is: -23 ln (1-R(1-exp-100/23)),
where ‘ln’ is the natural logarithm; ‘exp’ is the exponential transformation;
'R' is the rank value; '23' the scaling constant to minimise cancellation
effects. The indicators are combined using the domain weights shown in the
weighting section below. An equal 50% weighting was applied to the housing stock
and household circumstances domains, and then all indicators within those
domains were given equal weighting. This is done separately for each country.
Using the Python package Selenium, we scraped the directories of installers of
these websites and a risk assessment was conducted to ensure there were no GDPR
or other data concerns. For sites like TrustMark, Checkatrade and TrustATrader
which have many different types of tradespeople, we used search terms related to
insulation to obtain a list of insulation installers. This was combined with the
listed installers from NIA, IAA and Which? Trusted Traders (this was not
web-scraped as we had direct access to the list). The list was then fuzzy
matched by name and address to details from Companies House and de-duplicated to
avoid double counting across the websites. Fuzzy matching was done as there were
some companies which had similar but not identical names across the websites (eg
XY LTD and XY LIMITED). This method allowed us to identify that these companies
were the same based on the high percentage of similarity. We then added a marker
to identify whether a trader is a certified installer. The most common
certification for insulation installers is Trustmark, as it is necessary for a
supplier to have this to be able to carry out work funded by a government or ECO
grant. However, installers operating only in more affluent areas may have less
incentive to get this certification, since little of their work will be through
government schemes. Therefore, we include two additional certification schemes
to indicate if an installer is properly certified, the IAA and Which? Trusted
Traders. Installers belonging to any of the three schemes were classified as
certified. It is likely an installer will be willing to travel to perform a job
and this may cross over to other LSOAs/constituencies/local authorities. We have
made an assumption that they are willing to travel a radius of 25 miles from
where their offices are located. The 25 miles is equivalent to a journey time of
approximately 25 to 45 minutes (with installers driving time based on average
driving speeds in the UK for motorways and ‘A’ roads), which was considered a
plausible assumption on how far installers are willing to drive for a job based
on the expertise of our Which? Trusted Trader team. By mapping out a radius of
25 miles from the offices of each installer, using geospatial Python packages
Geopandas and Shapely, we were able to ascertain their coverage area. These were
then transformed into the number of installers available to service each local
area. The numbers of all and certified insulation installers per local area were
standardised by calculating the per capita figure (dividing the number of
installers over the total population of the LSOA/constituency/local authority)
and multiplying it by a 1,000 people for LSOAs and 100,000 people for
constituencies and local authorities. Each local area was then ranked using this
standardised number.
Weighting:
Domain Indicator Weighting Housing stock Wall insulation 10% Roof insulation 10%
Floor insulation 10% Property type 10% Property age 10% Household circumstances
Fuel Poverty / Income 16.7% Health conditions 16.7% Age 16.7%
Notes on Representation:
The data consists of ranks on a country-by country basis, leading to reliable
comparisons between areas within countries. Comparison of the indicators across
countries is not recommended due to differences in the source data across
countries. The indicators in the local housing stock domain are all created from
energy performance certificate (EPC) data, which is available for all four
nations of the UK. Although EPC data represents by far the most extensive data
on the quality of local housing stock, it does not cover all local housing. Data
from ONS shows that EPCs represent around 64% of residential dwellings in
England and around 62% in Wales. The assumption is made that the conditions of
properties without EPCs will likely be similar to the properties with EPCs for
each postcode area, although it is recognised there may be an unobservable bias
in the data, for example houses with EPCs may be newer or have better insulation
on average.
Notes on Bias:
Separate to the index but as a part of this dataset is the supply of insulation
installers which was web-scraped. As this was done online, there could be
tradespeople without an online presence who are not included. However, if they
are properly certified, they will likely be included to their inclusion on the
sites we scraped.
Access and Governance
Usage
Resource Creator:
- Which?
- Maeve Murphy Quinlan (https://orcid.org/0000-0003-2958-1008)
- Francesca Pontin (https://orcid.org/0000-0002-7143-8718)
Is Referenced By:
Data Use Requirements:
NO REQUIREMENTS
Access
Jurisdiction:
Great Britain
Data Controller:
Healthy and Sustainable Places Data Service
Availability Status:
Active
Licence:
UK Open Government Licence V3 (OGL)
Format and Standards
Language:
English
File Format:
- .csv
- .pdf
- .zip
Estimated Dataset Size:
<25MB