top of page

Building Age 


Detailed data relating to the age of cities' building stock is hypothesised to rapidly increase in value over the next several years. The primary driver for this is its growing importance in energy analysis. However until now remarkably little attention has been paid to age data, other than by tax authorities and the conservation sector.


In 2007 construction date was identified by Schiller (2007) in his assessment of German stock) as an important component for energy efficiency calculations, Arsoezen et al in their 2014 study into the significance of specific building attributes for 20,000 Swiss buildings in determining energy consumption also concluded that construction age was a crucial parameter. They found that buildings constructed after 1980, and before 1921, had below average energy consumption whereas those constructed between 1947 and 1979 averaged highest


. This finding is particularly relevant in the context of figures which suggests that the building and infrastructure stocks of most European countries have doubled since 1945 (Kohler, Steadman, Hassler 2009). Tanikawa and Hashimoto Tanikawa (2012) show age data to be essential in analysing long term stock survival rates and predicting demolition, while Stanilov’s work (2012) demonstrates its significance in predictive analysis for new build. Other examples of its varied use include its assessment of vacancy rates to identify industrial properties with residential potential to minimise greenfield land development (ODPM 2004) and in the analysis of health risks in dwellings (Haaselaar’s 2009)38 37 In the EPIE’s European study (p5) date of stock a characteristic assessed. In the residential sector, the age of a building is likely to be strongly linked to the level of energy use for the majority of buildings that have not undergone renovation to improve energy performance.38 Date data was accessed through the Netherlands Kadaster (Email correspondence with Professor Ernst Haaselaar)Date is also significant in understanding stock diversity and capital value and, it is argued, the attraction of intellectual capital. Research undertaken as part of the MRes course39 identified a close correlation between high value stock in London and that built before 1840 (listed buildings in blue) and largely before 1919 (conservation areas in yellow) shown in the images below, This is reinforced by Huntley Hooper’s findings that high end period property is more likely to achieve and retain premium values than luxury new build. Figure 17: 2013 London House Prices40. Figure 18: Designated Assets English Heritage 2012 39 Essay submitted by Polly Hudson for the MRes Knowledge Power Module40 Despite a growth in awareness of the importance of Jane Jacobs’ early insights into the value that older stock has in providing a spectrum of spaces and rent levels, little attention has been given either to measuring diversity levels by date or in modelling the economic and social impact of different buildings’ loss. 3.1International and historical context The first citywide colour-coded, stock date maps at footprint level were produced in Vienna by Hugo Hassinger in 1916 (Whitehand 2007). Hassinger was the first scientist to show the present state of a city’s stock and its genetic growth. His unusual use of colour to illustrate character and date was noted at the time (though interestingly recent buildings were left uncoloured), His work formed part of the emerging and important field of urban morphology (spearheaded in Britain by Michael Conzen after the war) and his Figure 19: Art Historical Atlas of Vienna. 191641 41 Photographed courtesy of the British LibraryInterest lay in the power cartographic visualisation to show coherence within a region42. No other citywide date maps have been identified from the past century other than those released online since 2013 (discussed on p.20) when Thomas Rhiel43 produced his stockdate map of Brooklyn. Though these were all made possible through the availability of new technologies and open data, motivation for their production differed. Rhiel, a political science graduate and web developer saw the maps as a tool to facilitate in depth Figure 20: Brooklyn stockdate map. Thomas Rhiel discussions on the political and socio-economic context of his local area. Email correspondence with Brendon Liu and Bert Spaan, developers of the New York and Netherlands maps respectively, shows their interest to be more in experimenting with Rhiel’s approach and in visualising and freely disseminating novel city metadata, than in the value of the data themselves44. The online reaction and rapid subsequent 42 . This also involved the creation of a topographical index of Austrian monuments to be created and a rare link between science and the emerging conservation movement. Hassinger went on to use genetic mapping to justify expulsion of non Germans during WW11.43 44. Though Spaag referred to possible use of data in the Netherland for energy analysis neither knew or were following specific ways in which the data was being used.development by other US cities demonstrates, and the level of public interest demonstrates the huge potential role both data and developers have to play. 3.2Digital sources (restricted) The Valuation Office Agency The VOA’s ‘Property Details’ database for dwellings was introduced in the 1970s and contains sixteen attribute classifications including building type, area, number of storeys and construction date. Data collected pre 1970 is considered to be less reliable. Since 2004, many of these records have been enhanced by inspections and third party data. VOA stockdate data is available for downloaded via the London Datastore, but only aggregated at LSOA level. (Figure 20). Though the VOA has shown interest this study it has confirmed it cannot release details at record level45. Eleven classifications are given for age beginning with ‘Pre 1900’ stock and continuing to the present with intervals of 20, 10 or 8 years. For commercial and industrial floor space date bands are Pre 1940, 1940-70, 1971-80, 1981-90, 1991-2000 and then annually from 2001 to 2003. No data was collected between 1939 and 1945. According to VOA significant number of properties are also classified with an unknown age coding, as it is/has been impossible to gain access to all dwellings and provide an accurate assessment of age; other information being derived from plans. 45 Correspondence with Paul Collins, VOA 2014/2015 with UCL’s Energy Department has also had its request for date data declined.. Figure 21: VOA date data visualised at MSOA level (London Datastore) 46 Figure 22: Distribution of pre 1919 buildings at LSOA level Maximum granularity of VOA releases 46 English Housing Survey Aggregated figures for dwelling date are also released by the English Housing Survey (EHS)47. EHS provides aggregated data on the English domestic stock, with Scotland and Wales undertaking their own housing condition surveys. Date data used in the UK’s Domestic Carbon Model (EIO), and in many government analyses derives from this source. Data is collected via interview with 13,300 households per annum and a physical inspection of around 6,200 of these homes. The former relies on owners’ knowledge of construction date. This is considered is unreliable. Commercial providers The GeoInformationGroup sells date data as part of its ‘UKBuildings’ products line48 The company claims to have detailed attribute data for 14.3 million properties although it unclear as to the proportion that are actually dated . Building age and type classification are given as having ‘90%+ accuracy levels (for 68% confidence limits)’. Commercial data is understood to be provided by bespoke ground surveys49. Costs are awaited. Figure 23: GeoInformationGroup UKBuildings 47 participating-households. The EHS is a continuous national survey commissioned by the Department for local government and communities (DCLG) collecting information about housing circumstances and the condition and energy efficiency of housing in England48!ukbuildings/c1his49 UCL Energy Institute3.3Conservation sector Involvement Collaboration with the conservation sector is seen as critical for the development of accurate stockdate data. Engagement with the conservation community’s extensive network offers an opportunity for access to significant dating knowledge and expertise A framework for the digital characterisation of the UK landscape already exists within Historic England’s pioneering Historic Landscape Characterisation (HLC) programme. Set up in the 1990s to provide a more informed starting point for advising on sustainable management, HLC’s highly ambitious plan, to use GIS to characterise all land in England and Wales, is almost complete. Quality in HLC maps however differs between local authorities, with urban areas carried out with a relatively broad brush approach50. HLC’s methodology developed out of the Conzenian tradition, in its recording of the evolutionary development of land through colour coded maps, according to land use and historical period. It gives equal prominence to all land types and its approach is an important influence on this study. Historic England and The Survey of London, UCL and the Pevsner Guides have all been consulted regarding the dating methodology. 3.4Data sources The Pevsner Guides The Pevsner Guides represent most comprehensive architectural survey of England’s buildings, and are based with, and published by Yale University Press. The guides were created by the German born scholar and conservationist Nikolaus Pevsner, between 1951 and 1974 (with his collaborators) and are updated on a rotating basis. Pevsner represents 50 Email correspondence with Roger Thomas at Historic Englandthe most scholarly analysis available on the date of London’s current stock51, providing almost a street by street account and covering an estimated 60% of Camden’s buildings of all types and ages., Conservation Area Appraisals Conservation Area appraisals are detailed sources of local information, often with historical maps included. These are developed by local authority conservation departments with research outsourced often to local amenity societies. Over 50% of Camden is covered by conservation areas and many appraisals draw from Pevsner as well. Online Historical Mapping Access to comprehensive historical mapping collections that hold OS County Series and National Grid maps at 2,500 scale is required. Multiple survey dates can be accessed by universities via Edina Digimap’s outstanding Ancient Roam and Historic download Service. Local authorities hold collection within their GIS departments52. For equivalent access for pre 1980s maps, local archive departments provide the only free source. Paper version can be bought through the Godfrey series and digital map tiles from OS Landmark. The National Library of Scotland has been found to provide the largest online collection of large-scale OS historical maps, georeferenced53 though the selection for London is not comprehensive and insufficient for this study. The British Library also offers access 51 The Survey of London, UCL provides the most comprehensive account of London’s historical stock52 Local authorities were required to purchase digitised OS historical maps in the 1990s for historical contamination mapping53,to smaller scale OS maps,54 (Individual local authorities such as Bristol also provide online historical maps for their areas and project’s such as MOLA’s locating London past and the British Library georefencing programme are also useful). Specific independent map enthusiasts are beginning to purchase, scan and release historical maps, such as the Rocque and Agas maps of London, and upload them into the public domain55.It is hoped that the release of stock date maps of London will encourage this approach.. Other online sources Access to Google Streetview and Bing Birdseye is also essential for the work. Specialist websites run by local amenity societies and historic building societies are also relevant. Other sites, if known to be reliable, may also be used. 3.5Dating the buildings The aim is to produce a colour-coded visualisation at high spatial and temporal resolution, which allows stock age to be read at both the mirco and macro scales. A sample map covering the Camden study area has been produced using ArcGIS and OSMM. This took approximately 120 hours to complete included accessing sources and developing the dating methodology56.In the first instance a quick assessment was made of the building plan on OSMM, and its façade through Google Streetview. Bing Birdseye was used if the building was set back from the street. The Pevsner Guide was then consulted. Where no date in Pevsner 54,55,_1746#/media/File:John_Roc que%27s_Map_of_London,_1746.png56 Data collected in the 2010 Hudson/Jobst survey was found to be insufficiently precise for the study’s purpose.could be found, Camden Conservation area appraisals, the National Heritage List for England were checked. I nothing was found, online searches for information uploaded by national and local amenity societies or historic building trusts were made.For all remaining buildings Edina Ancient Roam historical OS Tiles were used to identify earliest and latest maps on which the building appeared. A visual assessment of the architectural style of the building was then made to provide a date estimate required for the DATE field.All buildings in time will be added to or altered in some way, from wings to small extensions and window replacements. Dates entered into the field relate as closely as possible to the time when construction began. Where complete replacement has occurred behind an original façade, the replacement date will be given, with an additional note entered in the DATE RANGE field57 Figure 24: UCL dated 57 Facadism, as it is known, is less common in Camden than in the CityThe dating method proposed has been specifically designed to be able to be undertaken, outside archive departments58 though contributions from archive researchers and historians are strongly encouraged. 3.6Data fields and data confidence Stockdate mapping requires the dater to be able to differentiate between architectural periods and between original and facsimile buildings. Involvement of the conservation sector, is however essential to maximise data accuracy which will increase flexibility in classification intervals. Data inaccuracy is most likely to occur where date intervals are derived from historical OS maps only, with the DATE field only an estimate. Accuracy indicators are built into attribute tables through the following three relevant fields:DATE: Each building is dated to a specific year to generate colour coding. If no specific date is known then an estimated date is given within the date range based on architectural style.DATE RANGE: A short date range indicates a high level of accuracy. A long span of up to 20 years means that construction date is only broadly estimated.SOURCE: Referencing of sources is critical as these must be able to be verified. This also shows reliability level of the source. Thousands of individual sources could eventually be included.This method allows for categories to be adjusted to suit different sectors’ needs. However the narrower the age band the higher the inaccuracy will be. Sources of data have been selected that provide information with a high degree of reliability. Community feedback methods such as those used by Bristol’s ‘know your place’ or Mapbox would in future be of value to allow the public to confirm accuracy, monitor change and add detail and depth59. A more flexible wiki type approach could also be considered.The importance of conservation sector involvement is demonstrated by Columbia University Preservation Unit’s evaluation of the New York City’s PLUTO open dataset (visualised by both Thomas Rhiel and Brandon Liu). 2,000 buildings in Brooklyn60 were assessed using block and lot files, old issues of the Real Estate Record and Builders’ Guide and historic insurance maps. Precise dates were found for about 1,500 buildings and narrow date ranges for another 400. Figure 25: New York stockdate map61 Figure 26: Brooklyn data accuracy levels 59 60 between PLUTO data, shown in brown, and Columbia’s finding in green can be seen in Figure 24. The Columbia evaluation demonstrates the difficulty required in accurately assessing date. It also shows the value of conservation specialists’ involvement, the caution with which EHS date data particularly should be treated, and the importance of comparison of date datasets62, 3.7Date Categories Many different date classifications appear to be in use in the dating of stock. Decadal linked classifications are considered to be the clearest and most relevant to the diverse way in which data may be used. 20 year intervals have been chosen for colour coding purposes. Stylistic periods are classified as follows: DatePeriodPre 1700Medieval to 17th century1700 - 1799Early-Mid Georgian1800 - 1839Late/Georgian/Regency1840 - 1859Early Victorian1860 - 1879Mid Victorian1880 - 1899Late Victorian1900 - 1919Edwardian1920 - 1939Interwar1940 - 1959WW111960 - 1979Post War1980 - 1999Late 20th century2000 - 2014Early 21st century Table 1: Date range and period63 62 UCL Energy Institute’s commission of commercial date data allows for comparison between datasets63 Reigns are rounded up to the closest date intervalComparison with VOA and EHS classifications is shown below. VOADate RangeVOA Date Interval (Years)EHS SubdivisionsNetherlands/KadasterStudy MethodologyPre 1900 -Pre1800 1800-18501850-19001760-791780-991800-191829-391840-591860-791880-991900-191920Pre 1919, 1919-19441900-19301900-19191920-192910 1920-19391930-193910 1930-1945 1940-1944 (Nodata)0 1940-591945-54101945-19641945-60 1955-6410 1960-751960-791965-7281965-1980 1973-82101981-19901975-851980-991983-199210Post 1990 1993-199910 1995-2005 2000-20088 >20052000-2015(though subdivided annually from 2010)>2008Unknown Table 2: Date classifications

bottom of page