Renewable energy in the United Kingdom/Rooftop Solar PV
Why? There are approx 800,000 rooftop solar PV installations in Britain, and together they provide around half of the country's solar capacity. If we can map them, at least their latitude and longitude, this information can be useful for tasks such as predicting how much solar energy we'll get on a particular day, or searching for new suitable rooftops for future installations.
Rooftop solar panels are often very visible - they should be fairly easy to map. And to a small extent, once mapped they serve as landmarks for within residential areas.
Solar panels mapped as areas are already displayed in OpenInfraMap or the data can be used for other analyses of the UK's power network.
How to find rooftop solar PV?
- Find them on foot
- spot them on the tops of buildings. Spotting them in real life is particularly helpful since aerial imagery may be too old or too fuzzy - see below.
- Find them in aerial imagery
- for this it helps to have good clear imagery, and also to flick back and forth between different imagery layers (e.g. Bing and Digital Globe). Note that often imagery is a few years old (especially Bing), and if the imagery is from 2011/12 it will miss out a lot of installations. This can still potentially be useful - the comparison between old and new imagery can help you to spot things. Take care to be sure what you're looking at - sometimes windows in rooves, or skylights, or other rooftop machinery can look a little like solar panel installations.
- Find them in street-level imagery
- such as Mapillary or OpenStreetCam - try the new Deriviste tool?
You might find that solar PV is more commonly to be found in particular places: social housing, as well as schools, libraries and other civic buildings.
Do remember to watch out for the difference between solar PV (photovoltaic), which converts sunlight to electrical potential, and solar thermal, which heats up water. Solar thermal systems often look like squares on the roof too, but often with pipes inside.
Types of rooftop solar PV
The vast majority of rooftop solar installations are of modular panels of solar cells, either mounted directly on the roof or (for flat roofs on a support lattice to angle the panel). However there are a number of minor technologies which may also be encountered:
- Solar tiles. Use generator:type=solar_photovoltaic_tiles. A group of houses with tiles has been mapped in Camrose Close, Bilborough, Nottingham, and can be seen on Mapillary.
- Solar slate. See this trade webpage for a little more information. No known installations mapped at present.
- Building Intergrated PV (BIPV). Appropriate tag not yet decided. See this trade page for an excellent introduction. Significant installations at Sandymoor, Runcorn, which have been mapped. The typical installation seems to be around 1kW with 4 250W panels which are mounted flush with the conventional roof tiles. Smaller installations may be hard to tell apart from typical solar hot water, but odd shaped layouts are more likely to be these.
How to tag rooftop solar PV
Both iD and JOSM have good presets - simply type "solar" in the preset search box and it's the only preset. The iD preset is complete enough without any added tags; for the JOSM preset you need to make sure to also select generator:method=photovoltaic and generator:type=solar_photovoltaic_panel. For full OSM tag details see Tag:generator:source=solar.
Mapping a rooftop's solar installation as a single node is fine, but it will be very helpful if you can map it as an area (it helps to estimate power output, and it also makes it much easier to be clear about exactly which set of panels got mapped). Some rooftop solar has been mapped by adding tags to the building, but this may cause issues with interpreting some tags for data consumers.
You don't need to map each panel separately, one object for the "array" should be sufficient.
In most cases you won't know the power output it gives, but it might be useful to record the number of modules, or the estimated area (see below). Both of these make it possible for data users to make a heuristic estimate of the power.
Optional extra tags:
- location=roof -- we recommend this to help discriminate rooftop from ground-mounted solar. (Standard OSM tagging doesn't mark a big difference, so this tag is useful.)
- generator:orientation=* -- which way do the solar panels point? (in degrees or 16 point compass directions) e.g. 180 would mean the panels are facing directly South. Knowing the orientation can help estimating its actual output. On some houses panels will be mounted on two (or more sides of the roof) in which case they should be mapped separately.
- start_date=* -- if you know it, tagging the start_date (at least to the year) will be very handy in future analysis, to know whether it should or should not be appearing in certain imagery.
- generator:output:electricity=* -- if you know the output of the array then add it e.g. "12 kW". If you don't know it, consider adding "yes".
A couple of extra tags that have been deemed not-too-important:
- generator:solar:modules=* -- the number of panels (please note this is not a standard tag) - a "module" is a single rectangular unit. Knowing this can help guess the power output.
- Estimated area: notional_area="* sq m" (please note this is not a standard tag) - typical modern modules are approximately 1.5 sq m.
The building itself... Since rooftop solar is generally on top of a building, it helps to have the building mapped in OSM too, of course! As an area, so that the solar item can sit within it. It may also help to have at least a partial address for the building including full postcode as this would enable power capacity per postcode to be calculated.
Statistics and existing data
The Ofgem feed-in-tariff data gives information at the level of postcode districts. See Dan's blog on the FiT data for visualisations and summary stats, as well as caveats about that data.
The 25 postcode districts with most panels are summarised below, these are a good place to get one's eye in (emboldened ones have been done - FSVO of done):
|PC District||Place||Installation||Notional Power (kW)||Actual Power (kW)||Pct of total|
|NG5||Rise Park & Bestwood Estates, Nottingham||1768||5855.91||5596.46||0.21%|
|LE4||Beaumont Leys & NW Leicester||1706||6163.23||5993.6||0.21%|
|LE3||Braunstone, Glenfield & W Leicester||1518||5000.55||4866.03||0.18%|
OSM has (as of late March 2019) about 10,000 items in GB tagged with generator:method=photovoltaic. Approx 3,800 of those were added by SK53 in Nottingham.
The Ofgem FiT data suggests there are approx 800,000 in the country.
Social Housing Providers
Early investigations suggested that social housing providers - council at-length HMOs, Housing Associations and trusts, such as Peabody - were likely to be large-scale installers of rooftop solar. Economies of scale, common housing stock and tenant fuel poverty all have made these attractive projects. Using twitter SK53 asked a well-known historian of social housing to request examples from his followers. These are documented below:
- te Rose Hill, Oxford. Mix of interwar housing E part of estate and various ages on the W. Installation by Juju Solar.
- . 100 or so added on first pass. Whitehawk, Brighton.
- Fenham, Newcastle-upon-Tyne. An interwar estate with typical terraces of the period. Around 50 added.
- Stockport Homes. .
- Primrose Hill & Croftlands, Huddersfield. . A few mapped in Croftlands.
- Colchester Borough Homes . Some mapped in White City & Greenstead estates. These correspond to CO4 postal district (see list above).
- First Wessex Homes. 
- Bristol. All new social housing developments have solar panels 
- West Yorkshire, Incommunities. 
- Golden Gates Housing. Around 2000 installations around 2014-5. 
- Stirling 
- Perth & Kinross 
- Findhorn 
- . Panels only visible on high zoom levels of DigitalGlobe Premium and Standard Imagery, 125 added to date - all on new-build. King's Farm, Gravesend
- Edward Woods high-rise in Hammersmith & Fulham has panels mounted on walls . Don't know how to map/tag this.
Note that NRoSH data can be used to identify social housing areas in England.