Local Heat and Energy Efficiency Strategies (LHEES): phase 1 pilots - technical evaluation

2. Data requirements

2.1. Data sets utilised

We have provided a summary of the data sets used in Appendix A.

2.2. Commentary

It may seem obvious, but the right data is needed for the right purpose, so choice of data is critical. Access to the right quality data, at the appropriate level of detail, is essential for the delivery of a high quality and robust LHEES. However, collecting and maintaining data sets is a resource intensive (and hence costly) exercise and as part of the learnings from this assignment we have reflected on how we used each data source and the relative merits of each. This analysis is presented in Table 2-1 and should be helpful when considering which data sets are most worthy of further effort and investment.

From our experience in delivering the pilots, in selecting the data sets to be used we would recommend that a number of factors be considered:

• Policy and data
  • - Policy is a key driver in deciding what data is needed to support evaluation of opportunities, and subsequent monitoring and review. We undertook an initial policy review and prioritisation to guide the data sought.
• Expertise and approach
  • - Different local authorities have different levels of expertise and may wish to use different approaches. For example, only some have specialist expertise in data processing or capability for using geographic information systems. Nor is there consistency in preferred software to be used. In the pilot programme we generally attempted to deploy a standardised approach across the different authorities but this might not be the most appropriate decision going forward in all cases.
• How the data will be used
  • - Data can be better suited to some use cases than others. For instance, the generic information on building heat use is helpful in terms of identifying potential anchor loads for district heating networks but is not sufficiently accurate at the level of definition needed to actually determine the feasibility of a district heating network and significant further information would be required to actually progress the design of any scheme.
• Data quality and potential conflicts
• - In any data driven approach, the quality of the data is critical in terms of ensuring robust conclusions. There can be a perception that data that is generated and owned locally is of 'better' quality than national datasets but it is important to note that this is not always the case. Some of the data sets include a 'confidence' measure

The process highlighted a number of lessons that could improve the data in future:

- The data available for domestic properties is much more comprehensive than for non-domestic properties. This was a significant inhibitor in terms of preparing an LHEES for the non-domestic sectors and is discussed further in section 2.3

- In order for data to be used it requires the consent of those who control a particular data set not just to provide it which can be a time consuming process in terms of obtaining necessary consents. It is also important that the consents do not constrain the intended use (for example, being able to make the data public to justify conclusions / recommendations) and that the data is either provided in a suitable format or in a form that can be readily manipulated to fit the data structures employed.

- It would be worthwhile giving further thought to how data sharing could be facilitated to give data controllers the confidence to share data for these projects even when energy and climate change isn't the primary focus of the initial data collection

- The work highlighted the value of public sector data, and central collation of key datasets such as the Scotland Heat Map and Home Analytics.

- Some authorities highlighted issues where local data had not been collated into national data sets in time to allow them to be used in one of the Scotland wide datasets. Ideally data collection can be synchronised to minimise this risk.

- The Scottish public sector collects a wide range of data. This data has proved very valuable in revealing building energy opportunities through the LHEES work. It is important that the public sector continues to collect and maintain this data and expansion of the data collected (particularly in terms of the non-domestic properties) will improve the ability to deliver future LHEES.

- Local knowledge and data can be used to check and validate data and add greater depth. It can only really be used in the long term if it is turned into data that can be integrated with the baseline, preventing loss where someone leaves

A point to note, from both the Atkins supported pilots and others, is that the baseline data collation and manipulation into the right format to be used for subsequent analysis was quite time consuming. It is therefore important going forward that, where possible, standardised approaches are adopted to avoid the inefficiencies that would be created by each council developing their own bespoke methodologies and tools.

Data Set	What did we use it for?	Which councils / types of areas was it particularly useful for?	How important is it? How good was the quality / accuracy?
Scotland Heat Map	Used if individual council data sets were not provided for data needed as per below. Also used as the main property list if council data was not provided.	Where councils did not provide a full list of properties in study area with heat demand, architype, and age	If council data did not exist, vital to the study. This is a useful back-up to data held at local levels by the councils and was essential in the cases where particular councils did not have (or were unable to provide) equivalent data sets. For example, two councils provided full address lists, with another two providing address lists only for public or council owned buildings.
Heat Demand	In the fuel poverty analysis to estimate energy bills at Data Zone level, Also, in the EE and HDC analysis, heat demand allows us to identify potential DHN, and cost for new technologies.	For all councils	Very important – the bulk of our EE and HDC costing would not be able to take place without it. Similarly, with comparing the heating performance to the industry benchmarks. Confidence level 5 is taken from billing data, however the majority are lower confidence levels and are calculated using various assumptions and benchmarks. Having more billing data to contribute would make the heat demand, as a whole, more accurate. The challenge is that it constantly needs updating as things change.
EPC data	Not used separately, but is partly used within the domestic heat demand dataset. EPC data is almost exclusively domestic, as many non-domestic buildings are not currently required to provide and EPC.		Home Analytics includes the EPC dataset. However, it is noted that it doesn't give a separate estimate of annual heat demand. Instead it gives total energy demand which includes space and water heating but also electricity for lighting and ventilation (pumps, fans). EPC data in the heat map includes an estimate of annual heat demand. It is also important to note that Home Analytics only covers the domestic sector.
Information from councils	(As part of the BDR data) Identifying opportunities for domestic energy efficiency (wall and loft insulation) and low carbon heating upgrades (solar thermal and heat pumps).
Housing stock	Address list, tenure, sometimes additional information like architype, age, construction, EE measures etc – councils hold different information	For all councils	Adds to tenure information which is needed for the weighted spatial analysis. Properties more favourable for works if owned by the council
Local development plan(s)	Local development plan.	For all councils.	Can be helpful to know where new developments are coming, might be able to install decarbonised heat options, rather than retrofitting – not essential for analysis but useful for decision making and understanding potential future loads.
Project register	Energy consumption information and EE measures.	Very rarely supplied	Varies council to council. Home Analytics data used, for domestic properties, when this did not exist. A lot of the data was based on areas and not properties, and could not be tied to the property without a UPRN. If it did have UPRN attached, then it gave information on EE measures and energy consumption for a property.
Assessor domestic	Property age, property architype, floor area.	For all councils	These are essential to be able to identify opportunities and compare heat usage to benchmarks.
Assessor non-domestic	Property age, property architype, floor area.	For all councils	These are essential to be able to identify opportunities and compare heat usage to benchmarks.
National datasets
Fuel poverty information	A number of data sources were used. Primarily the Changeworks fuel poverty map together with Data Zone level SIMD income deprivation and Council Tax Banding A-C for ranking areas suitable for domestic area-based granting schemes (i.e. HEEPS:ABS).	For all councils	A single approach to this could be helpful. The Changeworks fuel poverty map is dated as it was created a few years ago. Particularly in areas where fuel prices have come down the last years (i.e. with many households on oil) the rates will be a bit lower. Nonetheless, the relative differences between the areas make it useful to rank Data Zones for recommending domestic area-based granting schemes (i.e. HEEPS:ABS), and will change with the new fuel poverty definition (this is highlighted in the fuel poverty analysis report). There is currently no housing costs data available to update the fuel poverty map to the new definition.
Home Analytics	Identifying opportunities for domestic energy efficiency (wall and loft insulation) and low carbon heating upgrades (solar thermal and heat pumps). In the fuel poverty analysis, the EE-bands were used to do a rapid assessment of how many properties with no identified upgrade that have an EE-band of D or lower. Also includes a modelled estimate of fuel poverty likelihood.	For all councils	This dataset is a collation of several other datasets. Accuracy depends on the amount of data from the feeder datasets such as HEED and EPCs that the address level data is based on. We identified some errors between address level data in Home Analytics and corresponding EPC certificates (internal brick walls on EPCs ended up in Home Analytics as external walls, rather than cavity walls). EST have been informed and mentioned their methodology will be adjusted. This was not a systematic check though. Vital when councils did not provide any information. Most of the information would be blank without this dataset. The Home Analytics is a useful coordination of several domestic datasets. However, as a combined dataset this does constrain some use of the data, for example to ensure it is not combined with another dataset in a way that double counts data. Having the separated feeder data provided in addition, as with the Scotland heat map would allow more flexibility with the use of this data.
Scottish Federation of Housing Associations	Tenure	For all councils	Additional tenure information if not supplied by council. Information only partial. Came from original heat map dataset. Does not include Council Housing.
National BEIS data	In the fuel poverty analysis to estimate energy bills at Data Zone level, together with heat demand data.	For councils for which a fuel poverty analysis is undertaken	The good thing about BEIS data is that it is real consumption rather than modelled energy consumption, which we know can differ substantially. The electricity consumption per meter, however, should not be confused with household level consumption as some properties will have more than one meter. An adjustment of the BEIS data to contain average consumption per household would be a major improvement. The BEIS data is already used to scale the domestic heat demand estimates in the Scotland heat map. It is noted that the BEIS consumption data only covers gas and electricity (although these are the dominant fuel types for most of the pilot areas). It is also only available down to a data zone level rather than for individual buildings.
Non-domestic EPC	EPC rating, age, fuel and architype where it didn't exist elsewhere.	For all councils	Most of the time non-domestic EPC data did not exist anywhere else
Listed buildings Scotland	Constraints	For some councils	Highlight properties that may be constrained on works to be done
Conservation areas Scotland	Constraints	For some councils	Highlight properties that may be constrained on works to be done
Gardens and designated landscapes	Constraints	For some councils	Highlight properties that may be constrained on works to be done
Scheduled monuments	Constraints	For some councils	Highlight properties that may be constrained on works to be done
Ancient woodland	Constraints	For some councils	Highlight properties that may be constrained on works to be done
Census 2011 data	Using the Census geo-demographic area classifications for raking Data Zones that are more likely to be 'able to pay' than others.	For all councils	This data is domestic only. It is getting more dated per year, as the Census is only repeated every ten years and the last one stems from 2011. Census data does hold information on heating type. Alternatively, Mosaic data from Experian could be purchased (though some of the relative Mosaic data is based on census data), with the drawback of it being another cost to the project. Used for sense check only.
Scottish Index of Multiple Deprivation	Using the SIMD income deprivation (together with fuel poverty and Council Tax Banding A-C at Data Zone level) for ranking areas suitable for domestic area-based granting schemes (i.e. HEEPS:ABS)	For all councils	From the SIMD, the income domain, rather than the overall index, is the more useful one recommending areas for area-based schemes. The housing domain of the SIMD is currently not very relevant to LHEES production, as it is based on overcrowding and central heating data, not energy efficiency. Unable to feed in to weighted spatial unless another indicator of deprivation is sourced.
Special protection area	Constraints	For some councils	Highlight properties that may be constrained on works to be done.
District heat networks	To identify if there were current DHN.	For some councils	Important when looking at other district heating options, options to consider expanding the current networks or improving on them. This came from the Scotland Heat Map or local authority.
Environmentally sensitive areas	Constraints	For some councils	Highlight properties that may be constrained on works to be done.
Site of special scientific interest	Constraints	For some councils	Highlight properties that may be constrained on works to be done.
Special areas of conservation	Constraints	For some councils	Highlight properties that may be constrained on works to be done.
Council tax data at Data Zone level	Using Council Tax Banding A-C at Data Zone level (together with the SIMD income deprivation and fuel poverty data) for ranking areas suitable for domestic area-based granting schemes (i.e. HEEPS:ABS)	For all councils	Alternatively, council tax data from the council can be used, or summarized address level data from Home Analytics. This is the underlying data for the Scotland Heat Map. Data is provided by Assessors to Councils. The Council tax band is used as a proxy to target different housing types by some Councils.
Gross household income distribution estimates at Data Zone level	The spread in weekly income at Data Zone level, together with the estimated energy bills and projected savings through energy efficiency installs, was used in the fuel poverty analysis to estimate potential fuel poverty alleviation through installing domestic upgrades.	For councils for which a fuel poverty analysis is undertaken	Gross household income distribution estimates at Data Zone level are dated as they were created in 2014. However, they are is the most recent publicly available income estimates at Data Zone level, so favoured over income related data from the Census 2011.

Table 2-1 - Data set commentary

2.3. Data for non-domestic properties

In general, there is better national data for domestic properties than non-domestic. This reflects ongoing programmes to support energy measures in the domestic market.

In general, there is relatively good data for domestic properties. However, for non-domestic properties, it was a general finding of the pilot that there was insufficient data to enable a robust LHEES to be developed for the non-domestic sector. In particular, the following issues were common across all the local authorities in the pilot:

• It was not possible to quantify the carbon emissions that the non-domestic properties were responsible for (and hence not possible to quantify the savings) as fuel type for non-domestic properties is largely unavailable. Clearly, this data does exist within the energy supply companies. However, there are challenges as the data is by meter. This means there will be places where there is more than one meter in a building, and also more than one building supplied by one meter. An exercise of matching meters to Unique Property Reference Numbers would need to be undertaken. We would recommend that the potential for making this available to the local councils is investigated.
• Assessor data provided valuable information for the largest number of non-domestic buildings, although this was not as complete as the Assessor domestic data. Assessor data was not always available for every local authority or every building. A significant (30% to 60% depending on council area) portion of non-domestic properties did not have an allocated building type, generated as part of the Scotland heat map within the Assessor Non-domestic dataset. We understand this dataset is incomplete due to limits to the UPRN matches within the Assessor Non-Domestic data when it was first incorporated into the Scotland heat map. The lack of the Assessor Non-domestic data made identification of potential measures more difficult.

The ideal outcome would be a range of data that provided more complete information on non-domestic buildings similar to that available for domestic buildings. The public sector data was good and is managed and controlled by the public estate so maintaining and updating public data should not be resource intensive. The public sector data is valuable but only accounts for a small proportion of the total non-domestic buildings.

There are a number of options for improving business data:

• Updating the Assessor data within the Scotland heat map, assuming the UPRN match is improved.
• All buildings could be physically surveyed. This would require a significant resource. A process of updating the data regularly would be required as change of business is reasonably regular.
• In broad terms a refresh of data within the existing Scotland heat map would be useful where the data has changed or been improved sufficiently to justify this work. Clearly utilising existing data that has already been collected by the Scottish Government or its agencies or using its funding is the lowest cost option. Enabling this approach and supporting bodies to supply data and understand and manage liabilities will be critical to this approach.
• The Scottish government has now funded a further round of Local Heat and Energy Efficiency Strategy pilots. A number of the other pilots, beyond this one, have included or focussed on business data. Understanding the data used and the processes from these would likely help understand the options for the Scottish Government.

An approach investigating a range of data sources that together provide an improved cohesive dataset for business is a potential approach. Scotia Gas Networks provide data to the Energy Saving Trust for domestic energy modelling. As such Scotia Gas Networks could be approached to enquire if they could provide data to the Scottish Government on both domestic and non-domestic. This may also open up opportunities for the Scottish Government to talk with electricity utilities regarding provision of data which could also include key information such as peak load demand requirements for business that could be used to help differentiate demands /zones across Scotland.

We would also note the following additional points:

• Some local authorities collate business data. It is not known how wide spread this data is across Scotland. There is unlikely to be a standard approach to data collection. The data does reveal that a significant number of businesses are in domestic buildings, which has potential impacts on the different Energy Efficient Scotland programmes which could include targeting buildings that are both domestic and non-domestic.
• The Scottish Government fund a wide range of activities. Consideration should be given to collection of data linked to public funds, where this would be valuable in the energy transition. For example, the Energy Efficient Scotland area work could be used to collect data on building structure and use, as well as collating any improvements in a way that can be used in a Scotland wide dataset. Scottish Water licence companies who place warm / hot water into the sewer system, and this would be valuable information showing where potential opportunities for unused heated water that is currently a waste product. Resource Efficient Scotland run programmes that provide support for business around energy measures that would be ideal information for this work.
• The Scottish Government manages regulatory functions which also have the potential to provide data. Good examples of this would include: SEPA pollution regulatory functions, local authority for air quality and environmental regulatory roles, Local authority for development plan and planning function roles.
• Some national datasets do exist, but these are currently provided in area-based formats and not down to building level. For example, the NOMIS dataset which identifies business size (ie SME or other) and Standard industry Categorisation (SIC). NOMIS is mainly available at intermediate zone.
• Ordnance Survey/Address Base may provide some information on building use. However, we understand that only Assessor data currently provides floor areas.

The Scottish Government may give consideration for how data is collected and for what purpose, as this could enable more data to be easily shared for public benefit. Strong data systems and approaches will be needed to ensure confidence in data sharing. This links clearly to the Scottish Government stated approaches to public data and open data. It is noted however that this is not within the direct control of Scottish Government and compelling controllers to provide data may not be possible without specific legislation in that regard.

We note that a useful enhancement for the data sources would be to develop a confidence measure such that those using the data know its limits and the reliability of any conclusions drawn from it.

Making the base data open, or available under licence, is also a potentially good strategy that would allow others to explore options for using the data in different ways and coming up with new approaches that support the Scottish Government agenda.

For the councils participating in the pilot, the lack of data for non-domestic properties is perhaps the single biggest frustration and issue in preparing a LHEES. If the data sets for these properties could be improved this would very significantly improve the quality and robustness of the LHEES. This is a significant opportunity for the future.

However, we would note that in terms of achieving the aims of the LHEES (lowering carbon emissions and removing poor energy efficiency as a driver of fuel poverty) improving the data is only one aspect. Actionability of measures for the private non-domestic sector (which is the vast majority of non-domestic buildings) is an issue. Most properties are rented (with many being small businesses) and there is a lack of alignment between the cost of investments in improving energy efficiency (which would be borne by the landlord) and the benefits of that investment (which would be realised by the tenant). We would recommend that work be undertaken to determine how to improve the actionability of measures in parallel to any investment in improving the quality / completeness of the data.