Data

This page provides links to download data from the PBCC as well as extra information on licensing and methods for data users.

Data Access and Licence

All the bulk data within the PBCC is available to download under the Open Data Commons Attribution License.

You are free:

As long as you:

This licence does not apply to the website, vector tiles, or other hosted data which is only to be used via the official PBCC website. You must not link, embed, scrape or otherwise use the constituent parts of the website.

Source Code

The source code used in the analysis and to build the website is published under the GNU Affero General Public License v3.0.

Your are free:

As long as you:

The source code is published on GitHub

Citing the PBCC

If you reference the PBCC or use screenshots or extracts from the tool in publications please cite:

Morgan, Malcolm, Anable, Jillian, & Lucas, Karen. (2021). A place-based carbon calculator for England. Presented at the 29th Annual GIS Research UK Conference (GISRUK), Cardiff, Wales, UK (Online): Zenodo. http://doi.org/10.5281/zenodo.4665852

Bulk Data Download

LSOA Data – this is a zipped CSV file containing all the data for each LSOA in England.

Transit Stops – this is a zipped CSV containing all the data from the Transit Stops layer for Great Britain.

15-minute isochrones – a zipped GeoJSON file containing all the 15-minute isochrones for each LSOA in England.

PCT combined route network - a zipped GeoJSON file containing the combined route networks for commuting and travel to school based on the Propensity to Cycle Tool

Main Data Sources and Methods

This section highlights the main source datasets used and provide more detail about the calculation and analysis methods. Please note the dataset listed here are third party and may have different licencing restrictions.

Mid-year population estimates

To convert carbon emissions into average per-person emissions we used the ONS mid-year population estimates for LSOAs.

DEFRA Emissions factors

To convert energy or fuel consumption into carbon emissions we used the appropriate year of the DEFRA Emissions factors.

Small area household incomes

For several types of consumption emissions, we have estimated consumption based on average household income.

Gas Consumption

Annual Domestic Gas consumption per LSOA is published by the Government.

Electricity Consumption

Annual Domestic Electricity Consumption per LSOA is published by the Government.

Building Main Fuel Type

Building heating types come from the 2011 Census. Note that an alternative dataset from the EPC database is also provided.

Building Ages

The building ages are provided by the VOA via the CDRC data site. Note that the data is rounded to the nearest 10 buildings.

Other Heating Systems

For houses with heating systems not fuelled by gas or electricity we used the 2011 Census data on heating types. We then assumed that these dwellings had the same heating needs at the average gas heated home. We could then account for the difference in carbon intensity based on fuel type.

Energy Performance Certificates

We took an extract of the Energy Performance Certificate Database in 2020. Which was then cleaned and aggregated to LSOAs. We used the Code Point Open to match the addresses of properties with the right LSOA. However a small percentage of EPCs could not be matched with an LSOA and where excluded. The code used to clean and aggregate the EPC data is on GitHub.

Travel to Work Flow Data

To estimate distance travelled to work by different modes of transport we used the 2011 Census travel to work data which give the number of people traveling between LSOAs by mode of travel.

Travel to School Flow Data

The travel to school data come from the Propensity to Cycle Tool.

2018 Flights

The flights emissions data comes from several parts of the CREDS Long Distance Travel Project. We use the Civil Aviation Authority airports and punctuality datasets to get the number of flights, passengers, and passenger kilometres. These are combined with the DEFRA emissions factors to get the UK’s total aviation carbon footprint.

This method differs from the UK government methodology in several key ways. Firstly, we consider all flights not just commercial flights. This gives us a higher estimate of total passenger kilometres per year. Secondly, we attribute 66% of emissions of incoming and outgoing flights to the UK rather than the DTF approach of only considering emissions from outgoing flights. We do this because 66% of passengers into/out of UK airports are UK residents.

As PBCC is an England only tool rather than UK tool we exclude all international flights from Scotland, Wales, and Northern Ireland and halve the emissions form flights between England and the countries in the UK. This method likely slightly overestimates English emissions, for example a Scot flying to New York via London would only have the Scotland – London emissions attributed to Scotland, while the London to New York Emissions would be attributed to England.

The final difference is methodology is that we used the DEFRA emissions factors to convert passenger kilometres to kilograms of CO2 equivalent and included the effect of radiative forcing. While the DFT only considers direct emission of carbon dioxide. This increased the carbon footprint of flying significantly but we believe is more representative of the true impact of flying.

To attribute the England flight emissions to LSOA we used research on the Gini coefficient of flying in England by Milena Buchs and data on the average household incomes to apportion out the emissions the each LSOA. Finally these are divided by the resident population to give an emissions per person estimate.

km driven per year

To produce an annual estimate of kilometres driven by car and van each year we relied on two datasets. First the Motoring Along report provides the kilometres driven per LSOA in 2009, 2010, 2011. While the DFT holds this data for every year since 2001 they have not openly published it to date. So, we turned to the anonymised MOT records. The anonymised MOT results contain far less spatial precision, only having the first two letters of the postcode of the test centre, instead of the LSOA of the registered keeper of the vehicle. Note also that the locations that the vehicle is registered, the MOT test is performed, and where is parked overnight may not be the same. Nevertheless, the anonymised data aggregated to postcode districts give a region picture of how kilometres driven per year has changed.

Historical car registrations and average emissions

Data on historical car registrations and the average CO2 per km for each LSOA was provided by the Department for Transport Vehicle Statistics Team and was extracted from data collected by the DVLA and DVSA.

Public Transport Timetables

The bus and light rail timetables are published by Traveline while the heavy rail timetables are published by the Association of Train Operating Companies (ATOC). We used the UK2GTFS package to convert and clean the data into the GTFS format.

Travel time Isochrones

The travel time isochrones were produced with OpenTripPlanner and the OpenTripPlanner for R package. The travel times were calculated for each LSOA’s population weighted centroid. Some centroids were slightly moved to facilitate routing. For example, if the centroid is in the middle of a lake then routing would fail. To build the OpenTripPlanner graph the OpenStreetMap was used as the road network extracted by Geofabrik terrain data was extracted from OS Terrain 50.

LSOA boundaries and population weighted Centroids

The boundaries and population weighted centroids for the LSOAs are published by the UK Data Service.

ONS Area Classification

The Office for National Statistics produce the Area Classifications.

2018 Consumption based statistics

The UK consumption based carbon footprint is used as the basis of the consumption part of the footprints. The overall UK footprint is scaled to account for England’s population and then distributed to the LSOA’s based on household income.

Ordnance Survey Open ZoomStack

The base map is a modified version of the OS Open ZoomStack.

Mapping Library and Tools

The web map was build using Mapbox GL JS and the vector tiles where built using Mapbox Tippecanoe. If you are interested in more information about building web maps for data visualisation see our tutorial.