|
dere-street.com
a variety of things, some useful. |
|
|
dere-street.com
a variety of things, some useful. |
|
Edinburgh was chosen as the focus of this study for several reasons. As the capital of Scotland, it occupies an important role in the urban hierarchy of the United Kingdom. Popularly known as the ‘Athens of the North’, it is famous for its beauty, particularly the extensive historic centre with its combination of the Old Town, with Edinburgh Castle and Holyrood Palace joined by the historic Royal Mile, and the Victorian quarter to the north, with the famous Princes Street and its gardens, with the New Town stretching to the North with its views towards the Firth of Forth and beyond to Fife.
In addition, Edinburgh has considerable cultural significance, as the Scottish capital and as a noted international cultural centre, due to its historic role and situation, complemented by the numerous international festivals it hosts every year. It has become a significant tourist centre, complimenting its cultural and historic roles.
A significant part of Edinburgh’s appeal is the lack of large scale organised urban redevelopment in its central core. This means that much of the historic fabric for which the city is justly famous has been preserved, with the only significant exception being the considerable amounts of 60s and 70s development along Princes Street. In particular, this means that the development of these areas has not been significantly shaped by the presence of large local malls, as the two small malls in the Princes Street study Region (there are none in the St. Giles Region) are both located on the eastern edge/periphery of Princes Street, rather than being central to the main Princes Street/George St. axes. This means that we are able to focus more directly on streets themselves, rather than the internalised street of the mall.
A desire to focus on a relatively large urban centre in Scotland meant that we were realistically limited to either Edinburgh or Glasgow. Edinburgh was chosen primarily due to its ease of access, the numerous data sources, the possibility of multiple study sites, relatively convenient access to relevant materials and facilities, combined with a degree of familiarity with the study areas. It was also felt to be a more intrinsically interesting study area than was Glasgow.
Limiting the study area to Edinburgh means that it is not possible within the bounds of this study to make wider comparison between different jurisdictions and conurbations. While this may make any attempt to identify any particularly local nature of changes difficult, as there will not be a basis of comparison with a control area (for lack of a better term), this is more than offset by the ability to construct a precise and detailed chronology of local change. Developing a wide-scale and detailed comparison between different urban areas was felt in any case to be beyond the scope of this project.
The types and quality of available historical data both constrain and define what types of analyses can be retroactively performed. In this case, when attempting to study the changing street, it was practical to focus on data sources that could be used to construct a picture of change and the types of change on Edinburgh’s streets. As a result, the sources discussed below are overwhelmingly biased towards recent and historical/archival records of building and site change, as it was felt that these provided the best means to reconstruct a picture of the changing city street in Edinburgh. Other, more qualitative methods of data collection were not used as primary data sources as, though they are both important and significant in their own right, they were not felt to be an appropriate basis on which to build the analysis, given that one of the research aims is to develop an understanding of urban change on the ground.
The Edinburgh and Leith Postal Directories were initially targeted as a sources of long-term historical data as they were easily accessible (they are available at a number of library sites and can occasionally be purchased from second-hand bookstores), seemed to be relatively complete, and they have the considerable added advantage that the information they contain is already organised geographically, by street address, which would make the identification of relevant records relatively straight forward.
In practice, however, these records did not provide the utility that was initially anticipated. The Edinburgh version of these Directories has not been produced since 1974, so they cannot provide any data for changes within the last twenty-five years. In addition, the Directories, while listing every owner/occupant of every floor of every building, do not explicitly identify either the ground floor occupant or the use of the site. While in some cases it was possible to identify the uses from the context of the entries, this was not generally possible. As a result of this, it was also not possible to gain information on site vacancies or any other related aspects of site change, like renovations or demolitions. While there is a substantial literature on the issues arising from the use of older (i.e. 19th century) versions of city directories , little information is available on the accuracy of their ‘modern’ incarnation as Post Office Directories. Given these issues, it was felt that the directories would be of limited use as a potential data source.
Thomson Directories, which are effectively the successors to the Post Office Directories, are available for Edinburgh from 1981, when they were introduced on a national level. As the data they contain was not organised geographically (unlike the Post Office Directories), it was felt that they were not an appropriate data source.
Goad Maps are a series of maps published by Chas. A. Goad for the use of planners and the retail industry. They detail the site level composition (i.e. the occupants and uses) of selected retail areas (including both city centre and out of town developments) for many of the larger metropolitan areas of the United Kingdom. Goad’s Edinburgh maps cover four central areas Regions withinof the city: Princes Street, St. Giles (the Royal Mile and south to Newington), Tollcross, Leith (primarily Leith Walk), and the Morningsidee area, which is some distance from the city centre. Taken together, these four central areas Regions cover broadly adjacent areas segments of the city which represent the majority of the city centre retailing (in the case of the Princes Street and Tollcross areasRegions, and to a lesser degree the St. Giles areaRegion) and two long radial/retail arteries connecting central Edinburgh to Leith in the north west (the Leith Walk areaRegion) and Marchmont to the south (contained in the southern extent of the St. Giles areaRegion). Copies of these maps are publicly accessible from the National Library of Scotland and the Central Library on George IV Bridge, and copies can also be purchased directly from Goad.
These maps are published approximately annually (though not to a strict schedule), and show plot boundaries, street lines, some basement sites, and their occupants and their uses. These maps are now computer generated by Goad, based upon site boundaries provided by the Ordnance Survey: prior to the early 90s they were hand drawn and lettered. Field surveys show the maps do not provide consistent and complete coverage of first and second storey uses. In most cases occupants are listed together with their usage, so it is theoretically possible to track occupant and usage change over time. The site data also reflect both changes in buildings themselves (inasmuch as site boundaries, buildings and street numbers change) and whether buildings are vacant, under construction, under demolition and so on.
Use of these maps as a primary source has problems, most directly in that we are limited to the geographic areas chosen by Goad: in the case of the St. Giles maps this means ignoring several interesting study areas, namely Victoria St. and the Grassmarket, while the Princes St. maps provide incomplete information for George St.. In addition, the maps have not been updated at regular intervals: for instance neither the St. Giles nor the Princes Street maps were updated in either 1986 or 1988, but both were done twice in 1987 and 1989. As is the case with the Postal Directories, the actual survey dates of the maps are also not made public, so it is not possible tell whether the maps are an accurate reflection of street use on a particular day or represent an aggregated data set collected over a longer period of time. This makes an analysis of change over regular time periods somewhat more difficult due to the degree of opacity inherent in using material which does not include specific published details on its collection methodology.
There is very little published information in the literature on the use of any of Goad’s data. Rowley provides the few references to this data source in the recent geographic literature, but his paper (and the subsequent discussion on its contents and conclusions in Area) focuses on Goad’s out of town shopping centre material. David McRobert Brooks’ unpublished Ph.D. thesis provides the only other use of Goad’s urban data that I have been able to find to date. As a result, there is little scope to comment on the relative completeness or comparative accuracy of the data source itself. This is perhaps surprising, given the considerable literature in North America on the use of Fire Insurance Maps as historical resources .
There is a long tradition within urban geography of using building records to understand urban change, usually through the urban morphology school. (For a detailed history of this school of geographical thought, see Whitehand .) With perhaps the obvious exception of J. R. Whitehand , this method of analysis is becoming less common in the literature. While it is possible to uncover substantial amounts of information by using these records, they do tell us little about how buildings are actually used after they are constructed, and to rely upon them as a primary source arguably infers that the role buildings play in their landscapes are fixed upon construction and do not change over time and in response to their evolving uses.
Two Ph.D. dissertations have attempted to use planning and property records as data sources, and while both found them to be technically available (in that there was no legal impediment to accessing the data), in practical and financial terms the data was inaccessible in the volumes the authors desired. Given that it was felt that the use of these records would not provide the most direct route to understanding many of the processes that were of interest, it was decided that they would be of limited use.
Copies of both commercial and residential tax records for the City of Edinburgh are available from the local council and through the public library system. These records are collected annually, but before the early 1980s commercial and residential uses were combined within the same dataset, which makes the separation of the relevant ground-floor occupants difficult. These records tell us who legally occupies a site (which may be different from the name the site trades under, a crucial point when trying to match this data with that provided by the Goad maps), who owns the site, and provides a contact address for all parties.
The Register of Sasines contains the official Scottish register of title deeds, which is open to public inspection. The data held by the Register is held on a plot by plot basis, and is currently in the midst of a long-term computerised transition to a mappable database known as the Scottish Land Information Service (ScotLIS), which will eventually be able to provide regional and local maps showing property ownership information. Due to the problems inherent in merging different databases accurately (see Chapter 4.1.7: Source Selection) it was decided not to use the Register as an additional data source.
There are several ground surveys of Edinburgh showing occupancy and use data, mostly provided for the Council as part of studies of parts of the central city, but they neither cover large relevant areas portions of the city nor are particularly frequent.
There were essentially two sets of data which could be used as primary data sources to construct some form of understanding of how the street is changing in Edinburgh: tax records and Goad maps. As the two data sources were compiled for radically different reasons, we might expect that they would place different emphases upon the data they present. The tax records provide a complete view of who owns and/or occupies every different site/business in the city, while the Goad data represents only data for street-level occupants and uses. The tax data is more detailed and far more complete in its coverage, but does not tell us some crucial information, particularly the name under which site occupiers are trading. This data is central to the Goad maps, but they, in turn, do not tell us who owns the property or which company (if any) occupies it. Thus there were trade-offs in using either as data sources.
When attempts were made to gain access to the tax records from the City of EdinburghEdinburgh Council it eventually became apparent that access to the records would not be easy and working with them not unproblematic. One year’s worth of data was collected by hand from publicly available library records and compiled into a database. During data entry it became clear that both the names and contact addresses for owners, agents and tenants were often wildly inconsistent throughout the data set, making it difficult to accurately compile a comprehensive list of the holdings of landowners and the occupancy patterns of individual tenants. This meant that it might be very difficult to develop comprehensive and consistent databases of owners, occupants and uses.
Comparisons of the results with contemporary Goad maps indicated the additional problem that there were not only numerous discrepancies between the street addresses of the tax records and those used by the Goad maps, but also between the definitions of properties and their uses by each data source. In many of these cases the tax records shows subdivisions of buildings which are not represented on the Goad maps, and as expected they also listed numerous non-street level uses not included by Goad.
Attempts to acquire electronic copies of portions of this data from the council were initially positively received but were eventually unsuccessful. While it was possible to continue to access microfiche copies of the records from the public library and input the data manually, it was quickly realised during the entry of the sample year’s data that this would be an exceptionally long and tedious process if a data set comparable to that provided by Goad was to be collected. Given the recognition that it would be time-consuming to manually transcribe the tax data, that the quality of information it contained was often variable, and that it would be difficult to work with and correlate with the other possible data source, it was recognised that the Goad maps, with their spatially organised data and their fixed geographical areasboundaries, would be by some degree the most suitable and effective data source to use.
Having decided upon a data source, it was still necessary to make decisions about which of the available data sets would/could be used. The Goad maps of Edinburgh provide a choice of four large areas Regions of the central city: Princes Street, St. Giles (The Royal Mile and south to Newington), Tollcross, and Leith (comprised primarily of Leith Walk and its immediate surroundings), and a more recent series of maps covering Morningside. While there is no doubt that the wider the area of study, the broader/more accurate the generalisations and eventual analysis, there also comes a point at which the sheer volume of potential data creates its own sets of problems. Given that there were four possible primary study areasRegions, and 17 years worth of pseudo-annual data for each, it became clear that there were essentially two approaches: either use annual data for a limited number of study areas Regions or broaden the geographical scope of the thesis and study all major potential study areas Regions but in somewhat less depth, by using data from every second or third year (etc.), an approach that had already been tried using some of Goad’s London data .
It was felt that data collected annually would provide the best opportunity to create the most accurate description of change from the available information. Given the constraints imposed by the sheer volume of information implied by using all of the potential study areaRegions, it was decided that the best way forward would be to limit the number of potential study areaRegions to two and to collect the data from each annually.
As a result, it was decided to focus on two primary study areaRegions: the St. Giles and the Princes Street areaRegions. These two areasRegions of were chosen for specific reasons, and with specific questions/analyses in mind. Given that the literature has identified numerous specific trends in urban change, (including increasinged consumerism ;, the manufacturinge of place and /space ;, the role of restructuring ;, the development of the so called tourist-historical city ; and, the retail revolution and so on) and their analyses of them (modernism, postmodernism and the like), and the numerous related analyses, as in the case of the Fordist/post-Fordist and restructuring debates, it was important to choose study arearegions that allowed for the exploration of the fundamental ideas underpinning these were as potentially representative and as relevant as possible/practicable to these debates, so that an appropriate empirical base analysesfor observation and comment could be constructed.
Princes Street was chosen as it represents a confluence of a number of relevant themes: particularly the retail revolution and the modernist/post-modernist analysis of urban change. This 1.2 km long strip has traditionally been recognised as Edinburgh’s retail and commercial core, and as such it would be expected to reflect the changes seen elsewhere in the United Kingdom: with the considerable levels of post-war site redevelopment it provides a fertile ground for analysis. The Region contains the actual shopping ‘High Street’ in Edinburgh (in contrast to the more tourist and leisure orientation of the ‘real’ (or more ‘literal’) High Street in the St. Giles study areaRegion) and as such provides a central focus for analyses of change within the prototypical retail ‘High Street’, while also containing considerable heritage and tourist interest.
The St. Giles area Region represents a quantitatively different perspective on many of the same issues. Its place in the retail structure of Edinburgh has been re-defined markedly in the last twenty years with the closure of several department stores, but the role it plays in the local economy, and iconography is, like Princes Street, central to what one might call the Edinburgh experience. Arguably this study areaRegion contains two neighbourhoods: that of the Royal Mile and its environs, and the more local shopping streets running south. The Royal Mile/High Street axis, running from Edinburgh Castle to the Palace of Holyrood, provides the core of the local heritage tourist trade. The St. Giles study areaRegion contains all of the built-up commercial development along this strip, which runs for 0.6km in an roughly East-West direction. In addition, the southern areas parts of the study areaRegion contains a significant amount of street retailing that supports local neighbourhood shopping, in contrast to the more nodal/hierarchical role of Princes Street.
Both areaRegions are now part of a changing Edinburgh, focused on consumption and the tourist experience. Both have been the focus of attempts to develop them in particular ways that have become typical: both have a systematically developed sense of place and identity, and have a focus on tourism and a more transient, space/place-centred consumption with associated leisure activities. They also offer the potential to discern different perspectives on more ‘local’ consumption practices, with the potential for comparison and contrast between the more ‘nodal’ Princes Street Region and the ‘local’ areas portions of the St. Giles Region.
Figure 4.1: Central Edinburgh showing the two study area Regions
Prior to 1978 the Goad maps for central Edinburgh (i.e. the Princes Street, St. Giles, Leith Walk and Tollcross) were produced approximately every other year: from then on, approximately annually. This change in publishing frequency represented a convenient starting point for data collection, and material from every map for the Princes Street and St. Giles study areaRegions published from 1978 to 1994 inclusive (17 years in all) was collated into two sets of databases, one for each study areaRegion. As can be seen from Table 4.1, there is a month’s difference in the dates of four of the seventeen Goad surveys for the two sites. For consistency, the survey date of each Princes St. map is referred to by the date of the contemporary St. Giles map for the four cases where the two are not identical, as the St. Giles data was the first of the two data sets to be transcribed.
| St. Giles Region | Princes St. Region | ||
| January 1978 | September 1987 | February 1978 | September 1987 |
| July 1979 | January 1989 | July 1979 | January 1989 |
| July 1980 | November 1989 | July 1980 | November 1989 |
| October 1981 | November 1990 | November 1981 | November 1990 |
| October 1982 | November 1991 | November 1982 | November 1991 |
| November 1983 | November 1992 | November 1983 | November 1992 |
| November 1984 | November 1993 | November 1984 | November 1993 |
| December 1985 | November 1994 | November 1985 | November 1994 |
| January 1987 | January 1987 | ||
Table 4.1: Goad Map survey dates 1978-94
Copies of the Goad maps from 1978 onwards were ordered directly from Goad, who provided copies of the maps for 1978-90 and 1993. Of all the maps that Goad provided, only the 1993 maps were original Goad maps: Goad did not provide direct or original copies of the other maps, instead supplying what appear to be copies of some form of archived materials. These maps were not direct copies of the originals, but were supplied as 51 mimeograph sheets (24 for the Princes Street maps and 27 for the St. Giles maps) copied by the British Library from what appear to be microfilmed originals. As a result the individual maps were spread over one, two or three unmatched sheets in what turned out to be a haphazard variety of scales and perspectives, and these had to be painstakingly collated. These maps were often of poor (and in some areas parts very poor) reproduction quality, and many suffered from considerable and irregular amounts of optical distortion. Some sections of some maps were not actually legible, and original copies in either the National Library of Scotland or the Edinburgh Room of the Central Library had to be repeatedly consulted as alternative sources for the missing data. Data for 1991-2 and 1994 were collected from copies of maps held in the Edinburgh Room.
The process of transcribing the information on the Goad maps was completed in two stages: the first was the transcription of basic site information for each year (site boundaries, street numbers etc.), followed by the transcription of the occupant and use data for each site. This process was then repeated until all the data had been transcribed.
The two original copies of maps provided by Goad, the 1993 maps for the St. Giles and Princes Street study areaRegions, were digitised into the MapInfo GIS program as the two base maps, as they did not suffer from optical distortion and were by some degree the most legible of the acquired sources. After all the buildings in each map bad been digitised, the basic site data (street names and numbers) was then added. These maps were then used as a basis for entering the site information for the other 16 years’ worth of data, with the help of a customised MapBasic application. Background information on the applications used and a technological overview is contained in Appendix 1, while the transcription process of this cartographic data itself is explained in much greater detail in Appendix 2.
Given that the internal database within MapInfo was not sophisticated enough to allow the construction of an appropriate interface to facilitate the entry of the occupant and user information, two programs were written to allow MapInfo to communicate with a FileMaker database and the occupant and use data for each site was then entered on a year by year basis. Specific details of the techniques used in entering this data are explained in detail in Appendix 3, while a more technical explanation of the programs that were developed is explained in Appendix 4.
During the transcription of the Goad data, two main data-entry issues became apparent: the first was the need to accurately link occupant and use information in the maps to the correct records in the database, and second was the need to accurately and consistently enter the data for each site.
Given that the data within the source material contained data organised visually (rather than in tables etc.), it was felt that some sort of visual/graphical input system would be appropriate to minimise input error and minimise any possible confusion in site identification. To this end, it was decided to transcribe the data into a GIS system, where the correct site could first be selected through a graphical interface/screen in a GIS system (by visual comparison with the addresses and site outlines in the printed material) before the actual site data entry would begian. This minimised the potential of site misidentification, which was felt to represent the primary opportunity for any significant data-misattribution.
Figure 4.2: MapInfo Map window
It was also recognised at the beginning of the data collection process that while the GIS system allowed the accurate identification of sites by organising selection within the mapping interface (cf. Figure 4.2), that the limited capabilities of the actual database within the MapInfo GIS program posed two problems: firstly, there were no error-checking or verification capabilities within the database that could be used to monitor and validate data entry, and secondly, that there were considerable opportunities for input error while manipulating information from within the table-style MapInfo database (illustrated in Figure 4.3), as the database itself provided little visual indication of which of a thousand possible records is actually being edited, allowing considerable potential confusion during the data-entry process.
Figure 4.3: MapInfo Database (‘Browser’) Window
While it was possible to create a data-entry template of sorts by custom-writing an interface to the database, it was felt that the more straightforward solution was to enter the data directly into a stand-alone database, which could be programmed to provide the necessary validation and error-checking, (see Figure 4.4) while allowing the maximum flexibility.
As the intention was to transfer the data to a stand-alone FileMaker database for processing and classification after data-entry was complete, this seemed to be the most straightforward approach.
As a result of this, to deal with these concerns about data integrity and input, a solution was developed where sites were initially selected for data entry within the MapInfo GIS program, the correct record was automatically selected in the FileMaker database, and then data entry was performed directly into the database program. As the GIS and FileMaker programs were not explicitly designed to be used together, it was necessary to write several additional software programs to allow them to communicate between each other.
Figure 4.5: Summary of occupant and use data entry process
The result of this development process is summarised in Figure 4.5. When the user selected a site in the GIS, the database was then instructed to select the appropriate record for the selected site, and after data entry was completed and verified, the GIS map was updated to show that data had been entered in the database for that site by colouring the site to indicate what data was now associated with it in the database. (Appendix 3 contains an extended discussion of this data-entry process.)
To minimise inaccuracies in entering site data, the layouts of the database were designed both to highlight the data for the current data-entry year, and to show all of the data that had been entered for the site (for every year) so the user could compare the data that was being entered with existing data for the surrounding years.
Figure 4.6: Ensuring consistent data entry in FileMaker with value lists
In addition, to facilitate the consistent entry of site use data, the data-entry mechanisms within the database were designed so that the user could chose to either enter the use data directly into the appropriate data field, or to select a site use directly from a pre-defined list of the 68 most common site uses. This allowed some degree of automation of the data-entry process and maximised consistency in data entry and classification. The entry options used are listed in Appendix 6.
The particular GIS/database system that was eventually developed and implemented was designed to ensure accuracy in targeting buildings for data entry and to facilitate fairly rapid and consistent data entry, and in this it was found to be highly successful.
Ideally, we would like our source data to be accurate, and frequently collected. In practical terms it is difficult to determine how accurate historical records actually are, especially given the lack of literature on the Goad records. While they appear to be accurate, and data entry reveals a remarkable internal consistency, there is still no mechanism, short of contemporaneous ground surveys, to check their accuracy. No attempts were made to corroborate the Goad data via a ground survey, as it was felt that this would be of minimal use in attempting to verify information that was almost two decades old. Observation did reveal however that basement and first floor usage was not consistently collected for all sites.
Some problems were encountered while using the Goad maps, although none of them were un-resolvable. The primary concern revolved around the very poor quality of many of the maps as supplied by the British Library for Goad, which meant that it was often necessary to check large amounts of information with original copies held elsewhere as the supplied copies were often either heavily smudged or otherwise illegible. In addition, severe optical distortion introduced during the copying process made the use of a number of the maps as digitising sources impossible. Data entry was also compromised to a degree by Goad’s failure to use a standardised and consistent list of abbreviations for the different site uses.
After data entry was complete, the initial data files were moved into a relational database and the database restructured. Records were copied to a relational database in which every record held one year’s site information (instead of each site’s 17 year history contained in one record) to facilitate the classification and analysis of the occupant and use data.
These new files were then checked to ensure that data entry had been consistent. The occupant records were then processed to give each site use a unique ID number (i.e. an occupant number). Sites which were vacant, under construction/alteration etc. were given a generic ID number. In an attempt to reduce the impact upon the database of occupant changes due to acquisitions, restructuring, mergers and so on, every case of occupant change was checked to see whether this change in the name of the occupant could be linked to some form of restructuring (for example when the Leicester Building Society became the Alliance and Leicester Building Society), and if so the database was adjusted to show this as not representing any change in site occupancy.
It was realised during data-entry that many of Goad’s usage classifications were somewhat arbitrary and often somewhat over-specific, resulting in classifications that were somewhat inaccurate or misleading (e.g. unisex hairdressers were often described as ladies hairdressers, clothing stores that catered mainly but not exclusively to women were classified as selling ladies wear only), and in many cases numerous classifications overlapped (e.g. pubs, taverns and bars). It was thus decided to develop a second, more abstracted usage description system that attempted to subsume many of the subjective use allocations within Goad’s data. During data entry, some similar use categories had already been somewhat rationalised: for example L & M/WR (Ladies and Men’s wear) and M & L/WR (Men’s and Ladies wear) were both entered as L & M/WR. This re-classification thus effectively extended this simplifiedcation of uses over throughoutthe whole of the database.
The result was the reduction of Goad’s 742 uses (covering both study areaRegions and all years) into 90 broader use categories. The impact of this re-classification is summarised in Table 4.2: the condensed list of uses used for the subsequent analyses are listed in Table 4.3.
| Number of sites with each use | Goad Use classification | Condensed Use classification |
| 1 | 160 | 22 |
| 2 to 5 | 57 | 22 |
| 5 to 10 | 19 | 17 |
| 10 to 20 | 7 | 10 |
| 20 to 50 | 3 | 8 |
| 50 + | 1 | 2 |
| Total number of uses: |
Table 4.2: Distribution of the number of St. Giles sites for a given use in the January 1978 dataset
It is clear from Table 4.2 that if we use Goad’s use data classifications, we must work with a dataset that has a preponderance of uses that exist only on a single site in the whole of the study areaRegion: in 1978 for the St. Giles Region these uses accounted for almost 65% of all of the reported uses in the Region. In comparison, the application of the condensed use classification (CUC) produces a substantially different profile of site counts: here only 27% of all sites were represented by single sites. This pattern is consistent throughout the re-distribution, with the reduction in numbers of uses having the effect of generating significantly larger numbers of uses with more than 5 sites. In this example year, the Goad data show only 12% of all uses occupy more than five sites, whereas in the CUC classification almost 46% of all the uses have 5 or more occurrences. In real terms, while the CUC classification reduces the overall number of distinct site uses by two thirds, the actual number of uses with 5 or more sites has actually increased, from 30 in the Goad use classification to 37.
In practice, this classification system meant that all pubs, bars and taverns (and so on) were classified as public houses, while all of the numerous variations of clothing-related stores were described simply as ‘clothing’, and so on. The two broad categories of ‘Clothing’ and ‘Home Furnishings & Household Goods’ by themselves originally contained some 179 of Goad’s original definitions, representing 24% of all the reported uses. The aggregation of Goad’s clothing-related use classifications is detailed in Table 4.4.
| Abstracted site uses (also referred to as the ‘Condensed Usage List’) |
||
| Amusement Arcade | Dwelling | Off Licence |
| Ancillary building - warehouse, loading dock etc. | Electronics/Appliances | Office Services |
| Antiques | Estate Agent | Offices |
| Artists materials | Film Developing | Optician |
| Audio Hi-Fi TV etc. | Financial Services | Outdoor & Sports Goods |
| Baker | Florist | P. H. |
| Bank | Food store | Pet Supplies |
| Betting Office | Foot WR | Post Office |
| Bingo Hall | Foot/WR Repair | Professional Services |
| Books etc. | Gallery | Record Store |
| Building Society | General Goods | Restaurant |
| C.T.N. | Gifts | Showroom |
| Cameras | Hair Care, Styling & Beauty | Solicitor etc. |
| Car Accessories, showrooms etc. | Hardware | Stationery |
| Car Park | Hall | Takeaway food |
| Cards | Home Furnishings | Theatre/cinema |
| Charity Shop | Hotel etc. | Tourist Services |
| Chemist | Institutional/Social Services | Toys |
| Church etc. | Jeweller | Travel Agent |
| Clothing | Knitwear & Tartans | Under Alteration |
| Commercial Services | Launderette | Under Construction |
| Computing etc. | Leather Goods | Under Demolition |
| Crafts | Medical services | Under Reconstruction |
| Demolished | Misc. Goods | University use |
| Department Store | Museum | Vacant |
| Derelict | Musical Instruments | Video Rental etc. |
| Dry Cleaning, Launderette etc. | Newsagent | Other. |
Table 4.3: Rationalised Goad Usage list
This level of abstraction was felt to be useful in laying the groundwork for later attempts to generate broader indexes of site change as more subjective changes in classification and of usage by site occupants would effectively be removed from the database.
| Goad clothing categories | ||
| Army Surplus | Haber. | Leather Wear |
| Bridal WR & Dress Hire | Haberdashery | Leather WR |
| Bridal/WR | Hosiery | Lingerie |
| C & L/WR | Jeans | LM/WR |
| C/WR | Knitwear | LM/WR & FT/WR |
| C/WR & Restaurant | L & C/WR | M & C/WR |
| C/WR & Soft Toys | L & M & C/WR | M & L/WR |
| Career WR & Uniforms | L & M/WR | M/L/WR |
| Casual Wear | L & M/WR & Foot WR | M/Underwear & Accessories |
| Clo | L & M/WR & Knit/WR | M/WR |
| Clo. & Fancy Gifts | L/Sep | M/WR & Leather WR |
| Clothes | L/Tailor | Millinery |
| Clothes Hire | L/WP | Nurs. & C/WR & Prams |
| Clothes, Toil, Ho/Goods | L/WR | Nylons |
| Clothing | L/WR & Bridal WR | Rain/WR |
| Clothing & Gifts | L/WR & Bridal/WR | Scout Shop |
| Dance/WR | L/WR & C/WR | Sheepskin & Leath/WR |
| Fab. | L/WR & Fabric | Shirts |
| Fabric | L/WR & Fabrics | Sweat shirt printing |
| Fabric/Sewing Machines | L/WR & Fancy Goods | T Shirt Print |
| Fabrics | L/WR & Tartans | Thermal WR |
| Fabrics & L/WR | L/WR/Gifts | Ties |
| Furs | Lea. WR | Ties & Scarves |
| General Outfitter | Leath/WR | Workwear |
Table 4.4: 72 Goad Categories condensed into ‘Clothing’
Frontage data for each site was calculated by first identifying the sides of each site which faced along the main street, and then calculating their length in the GIS program. This provides as fair a representation of the street level frontage for each site as can be calculated.
The values calculated for the each site’s area are, however, somewhat more ephemeral. As there is no distinction made in the Goad material regarding the uses within each site (i.e. what is floor space, warehouse space, unused/unusable space and so on) the area calculations more closely represent the physical size of each site. In addition, the accuracy of this data is to a degree affected by the inevitable errors that arise during the digitising process, by some inaccuracies in the Goad data, for example where upper or basement floors are not included in the original dataset, and from inaccuracies introduced by the need to add sites that did not appear on the digitised 1993 Goad maps by hand. It was not possible to either correct or compensate for these errors, but since it was empirically determined during subsequent analyses that calculations of rates of change based upon the site-area of changed sites, the street frontage of changed sites, or the numbers of changed sites with the appropriate data for unchanged sites revealed almost identical results. As a result, all further analyses calculated change (etc.) only in terms of percentages of sites.
In order to allow for a more local analysis of changes than was possible with analyses based upon the two study areaRegions, it was felt that the two study areaRegions should be subdivided into a number of smaller geographical areasunits. Sites were initially divided up according to what street blocks each was in, with each block representing the sites on one side of a street between two side streets.
Figure 4.7: Example Area and Block classifications, St. Giles
It was recognised that this division method produced a classification that was substantially flawed, as there were both large variations in the number of sites contained within each block (ranging from 8 to 40), and between the overall number of blocks across the two study Regions (22 in the Princes St. study Region, versus 49 within the St. Giles study Region). It was felt that as a result of this that working at individual block level was both too fine a scale to avoid the possibility of substantial changes in the composition of individual blocks resulting from the changing of relatively small (in real terms) number of sites, and simultaneously unwieldy, due to the large numbers of bocks across the two Regions.
In an effort to resolve these issues, larger (Area) units were created, typically representing agglomerations of neighbouring blocks in contiguous street Areas, often spanning both sides of the road. It was found that with some judicious allocations it was possible both to define a series of neighbourhood Areas that were roughly the same size and contained broadly similar numbers of sites. The differences between this classification and the block-level classification is summarised in Table 4.5.
| Count of sites | Block- based classification | Area-based Classification |
| 50 + | 0 | 6 |
| 40 to 50 | 0 | 4 |
| 30 to 40 | 2 | 5 |
| 20 to 30 | 10 | 1 |
| 10 to 20 | 25 | 0 |
| 1 to 10 | 13 | 0 |
| Total: |
Table 4.5: Effect of different site agglomerations in the St. Giles Region, 1978
Division of the Region’s sites into block areas has the effect of creating a substantial number of blocks with small or very small numbers of sites: in this example 76% are of less than 20 sites per block. By comparison, the grouping into larger Areas has resulted in only one Area containing less than 30 sites. With a number of the St. Giles Areas containing over 40 sites, it is clear that this form of site grouping ensures a more robust basis for any analysis of change or composition.
Details of the actual variations within each of these Areas is given in Table 4.6 and Table 4.7, while textual descriptions of the boundaries of each Area is given in Appendix 5.
Figure 4.8 shows the Area classifications of the St. Giles Region. In the majority of cases, the sites that were grouped together into Areas were along both sides of one or more streets. The only exception to this is Area 0. Given the significant geographic concentration of institutional sites within the centre of the St. Giles Region, in an area bounded by Market Street in the north, Potterow in the south, George IV Bridge on the west and Blair Street to the east, it was decided to group all of these buildings (and all their adjacent sites), which included the City Chambers (of the City of Edinburgh), the National Library of Scotland, the National Museum of Scotland, St. Giles Kirk, the Parliament buildings, numerous courts, municipal buildings and services, and several university buildings, including Old College and the Department of Architecture, into one single contiguous Area spanning the Royal Mile, the Cowgate, and Chambers Street.
| Area | Site Count | Total Frontage (m) | Total Area (sq m) | |||||||||
| Min/Max/Variation | Min/Max/Variation | Min/Max/Variation | ||||||||||
| 0 | 45 | 50 | 5 | 1,898 | 1,968 | 70 | 71,477 | 72,534 | 1,057 | |||
| 1 | 48 | 65 | 17 | 511 | 570 | 60 | 9,727 | 10,196 | 469 | |||
| 2 | 65 | 73 | 8 | 829 | 843 | 14 | 11,846 | 11,846 | 0 | |||
| 3 | 34 | 39 | 5 | 426 | 438 | 12 | 6,059 | 6,445 | 386 | |||
| 4 | 36 | 44 | 8 | 257 | 257 | 0 | 7,126 | 7,576 | 449 | |||
| 5 | 31 | 34 | 3 | 374 | 382 | 8 | 5,992 | 6,097 | 105 | |||
| 6 | 34 | 37 | 3 | 389 | 391 | 2 | 5,646 | 5,646 | 0 | |||
| 7 | 54 | 59 | 5 | 574 | 611 | 37 | 9,484 | 10,191 | 707 | |||
| 8 | 19 | 29 | 10 | 291 | 328 | 37 | 5,067 | 5,256 | 189 | |||
| 9 | 52 | 62 | 10 | 659 | 693 | 34 | 12,891 | 13,932 | 1,041 | |||
| 10 | 26 | 34 | 8 | 299 | 450 | 151 | 7,223 | 8,702 | 1,479 | |||
| 11 | 29 | 43 | 14 | 333 | 389 | 56 | 5,506 | 7,639 | 2,132 | |||
| 12 | 51 | 57 | 6 | 465 | 479 | 14 | 9,453 | 9,915 | 462 | |||
| 13 | 46 | 49 | 3 | 339 | 355 | 17 | 4,520 | 4,709 | 189 | |||
| 14 | 49 | 57 | 8 | 452 | 454 | 3 | 7,217 | 7,753 | 536 | |||
| 15 | 59 | 63 | 4 | 539 | 545 | 6 | 7,425 | 7,467 | 42 | |||
Table 4.6: Year to year variations within the St. Giles Areas, 1978-94
| Area | Site Count | Total Frontage (m) | Total Area (sq m) | |||||||||
| Min/Max/Variation | Min/Max/Variation | Min/Max/Variation | ||||||||||
| 1 | 57 | 63 | 6 | 712 | 715 | 3 | 17,114 | 17,220 | 105 | |||
| 2 | 61 | 77 | 16 | 842 | 928 | 86 | 14,047 | 17,141 | 3,094 | |||
| 3 | 60 | 64 | 4 | 696 | 696 | 0 | 15,570 | 15,570 | 0 | |||
| 4 | 61 | 64 | 3 | 615 | 624 | 9 | 6,910 | 6,935 | 26 | |||
| 5 | 17 | 55 | 38 | 305 | 611 | 306 | 5,196 | 9,096 | 3,900 | |||
| 6 | 39 | 48 | 9 | 403 | 462 | 58 | 11,910 | 12,515 | 605 | |||
| 7 | 77 | 85 | 8 | 500 | 516 | 16 | 6,949 | 7,161 | 212 | |||
| 8 | 60 | 72 | 12 | 719 | 785 | 66 | 17,515 | 18,538 | 1,024 | |||
| 9 | 61 | 65 | 4 | 889 | 892 | 3 | 20,364 | 20,364 | 0 | |||
| 10 | 56 | 60 | 4 | 791 | 795 | 4 | 19,724 | 19,770 | 46 | |||
| 11 | 34 | 39 | 5 | 448 | 464 | 16 | 9,971 | 10,166 | 195 | |||
Table 4.7: Year to year variations between the Princes Street Areas, 1978-94
Table 4.7 details the site data for the eleven Areas of the Princes Street Region. The expansion of the Goad dataset in 1984 is responsible for the significant variations in site numbers in Area 5, which contained the majority of the new sites, although there were also additions to Area 6. 1984 also marks the addition of Area 4 (The Waverly Centre) after its construction.
There are many approaches that can be taken in an attempt to understand both how the street is changing and, perhaps more importantly, what it is. Essentially, they can be broken down into two basic approaches: asking particular questions about how the street has changed in specific ways (e.g. has the number of leisure/recreational sites increased) and attempts to construct broader and more abstracted forms of analysis of the changing street, for example whether there is evidence to show increasing numbers of chains/multiples and whether this change is limited to one broad sector (e.g. retailing) or is evident across the wider range of site uses (e.g. retailing, services and leisure uses) and whether this has a particular geographic concentration.
Arguably, any attempt to analyse change will be arbitrary be to some degree: many of the indicators of change are by necessity inter-linked – i.e. for leisure uses to increase, other uses must change in response. Because of this, it is difficult (if not a logical impossibility) to simultaneously discuss change in detail while simultaneously analysing the broader patterns of change. The result of this is, of course, the recognition that the street is an unruly and recalcitrant subject, and that analysis at any level invariably calls for a counter-point from a different perspective, a different scale, to contextualise the discussions. This is particularly important if, as we would expect, different questions and different scales provide contradictory evidence.
The logical approach therefore would perhaps be to ask a series of specific questions about how the street has changed to provide a theoretical foundation before proceeding to the more general questions about the nature of the street and how it has changed. Following this approach, we can analyse a series of specific issues before attempting any wider synthesis of the results as a whole.
That streets change is both obvious and inevitable. However, defining change is perhaps less straightforward. This section will analyse the general nature of change and will focus on three basic forms of change: the changing street landscape, changing site occupants, and changing site uses. Subsequent sections will focus on more specific types of change, e.g. the growth of leisure, and the changing roles of chains and multiples.
Figure 4.10: Example of long-term site change, St. Giles
Changes in the built landscape were based upon an analysis of specific references in the Goad data to site-level redevelopment. The results are essentially comprised of the aggregation of all the site-change data that was specifically identified by Goad. Figure 4.10 provides an example of explicit site change within a section of the data for Area 6 of the St. Giles Region.
Analysis of the changing occupants in the study Areas relied upon the Condensed Use Classification of Goad’s occupant data that was described in Section 4.4.1: Occupant Identification. The occupant identifier data was then analysed to show how often site occupants changed, to identify sites which had large numbers of different occupants, and to determine whether different Areas exhibited greater amounts of change than others.
Analysis of changing uses relied heavily upon the re-classified Goad data whose creation was described in detail in Section 4.4.2: Use Classification. This data was analysed by Area to determine how many different uses there were, and to determine both how the types of uses were changing and if it was possible to identify any patterns of change in the types of uses in each Area and Region.
| Retail | Leisure | Services | Institutional | Other | |
| Shops | Restaurants, Pubs & Clubs | Banks etc. | Council uses | Vacant etc. | |
| Food stores | Museums, Galleries | Insurance & Finance | Churches | Entrances | |
| Clothing stores | Tourist facilities | Travel Agents | Universities | Car Parks | |
| Household goods | Hotels | Offices | Government |
Table 4.8: Broad Classification of site Uses
In addition to the use analysis described above, a more general use classification was constructed in an effort to identify broad trends in the changing sectoral makeup of the street that might be overlooked by the earlier use analysis. All sites were classified into one of five broad non-overlapping categories.
This data was used to contextualise discussions about the changing street by allowing reference to wider trends than focusing on single issues would allow. In particular, it recognises that within a geographically defined locale, for one general type of use to expand, the others will also be affected, and this analysis goes some way to quantifying the nature of these changes in Edinburgh.
Dominant conceptions of the postmodern street characterise it as a constantly changing, unstable, placeless entity. To focus exclusively on change ignores the tensions within the landscape between change and stability, between new sites and older occupants who may have been in situ for generations, that gives the street much of its character. In addition, focusing solely on change also implies that its opposite, stability, is simply its inverse.
This section questions the basic underlying assumption by focusing on how stable the street is, by focusing on how rarely some sites change, and whether we can develop an understanding of what might be described as the ‘relic’ landscapes that underpin the street. Thus, rather than focusing on how many and how often sites change, this section highlights both how many sites have continual occupancy and measures how ‘old’ the streetscapes of each street Area are in terms of the average age of the occupants it contains.
The role of the city and the street as postmodern leisure centres has become one of the central tenets of much of the current leisure studies literature. This section focuses on this assumption by studying the role and distribution of leisure services within the two study Regions in an attempt to understand both whether we can measure any appreciable growth in leisure services as a component of the overall street usage, and whether specific patterns of changing leisure uses emerges over the study period.
All site occupants and uses were divided into three broad categories: leisure sites, the places where leisure actually takes place, leisure facilitators, the sites and services that cater to leisure uses, and all other uses.
| Leisure Sites | Leisure Facilitators | Other | |
| Pubs, Clubs etc. | Travel Agencies | Non-tourist retail | |
| Restaurants, Cafes etc. | Bureaux de Change | Non-tourist institutional | |
| Museums and Galleries | Tourist Information etc. | Non-tourist services | |
| Heritage sites | Outdoor/Sports Goods etc. | Vacant sites etc. | |
| Tourist stores/services |
Table 4.9: Sample Leisure classifications
Defining tourist-related sites was by far the most problematic of these classifications. Services that cater directly to tourists, like hotels, and those that are leisure/educational attractions, like the National Museum of Scotland, are fairly straightforward to classify, whereas the wider aspects of the ‘heritage industry’ are somewhat more difficult to quantify. In these cases, the classification that was used relied in equal parts upon Goad’s classifications of souvenir stores, sites that sold ‘Scottish’ goods such as knitwear and tartans, whiskies, jewellery and so on, and upon informal field surveys.
Given that many stores and services serve both tourists and the local population (e.g. pubs, restaurants), it was felt best to only identify as tourist-related stores those that were primarily oriented towards tourists, so sites that could be said to serve both groups, like combination film processors/dry cleaners, were not identified as tourist-related sites. This decision was primarily made to maximise the consistency of classification of these retail/leisure sites across the wider study Regions and across the length of the study period: this has had the effect of possibly slightly underestimating the number of tourist sites, but there is no satisfactory means to resolve this dilemma, as field surveys of the contemporary occupants would still not allow us to apply anything other than a subjective correction for the sites and services that no longer exist.
The definition of ‘service’ sites was perhaps the most problematic of all the usage classifications. Two types of service uses were readily identified: ‘white collar’ services, which included medical and financial services, were easy to identify, as were the various service buildings, warehouses, loading bays and so on. Far more problematic were sites that provided a mixture of goods and services: chemists, dry-cleaners/photo processors and so on. Were these retail sites, service sites, or a combination of the two, and if so exactly what type of combination?
As it is possible to argue that the majority of retail and leisure sites are simultaneously providing some form of service provision, classification decisions revolved around the dilemma of definition. It became clear that it was not possible to generate any form of logical classification for retail sites, and it was decided to avoid such arbitrary distinctions by simply classifying all shops (etc.) as retailing. This has meant the service classification is based upon conventional definitions of ‘white-collar’ services. As the shift towards white-collar services is seen by many to be central to the new post-industrial, post-Fordist economy, it was felt that this was an acceptable trade-off.
All site uses were finally divided into three broad categories of services:
| Personal Services | Financial Services | Other | |
| Estate Agents | Banks, Building Societies | All shops | |
| Doctors, Lawyers etc. | Mortgage Services | Vacant sites etc. | |
| Travel Agents | Financial Services etc. | Ancillary/service buildings |
Table 4.10: Sample service classifications
The ‘services’ classification was divided into several sub-groups during data classification, purely to facilitate the classification of the relevant sites. Once classification was finished, all of the sub-classifications were re-aggregated back into a generic ‘service’ classification.
The final service classification provides a mechanism to focus on the provision of these services within the study Regions, and to see if the shift towards a service economy could be measured by studying changing street-level uses. It should be understood though, that it is quite possible that the results would not present a particularly accurate representation of changing service provision in the study Regions as a whole, given that the source material is geographically limited, there are difficulties in creating accurate and appropriate definitions of service uses, and perhaps most importantly, the fact that the source material does not provide any information about above-street site uses, which should in itself result in a substantial under-representation of the service activities within the study Regions.
While it is arguably appropriate to identify the street level as the primary location of retail (and to a slightly less degree leisure) provision, it is clear that the street itself does not monopolise service locations: to assume otherwise would, for example, ignore the factor of home-working. This problem is systematically compounded by Goad’s data collection/representation practices: in many cases the maps list ‘entrances’, even if they are shown to be attached to substantial plots of land, and without providing any indication whether these entrances service residential or other site uses. The result of this is that in many cases businesses and services that are clearly labelled as so at street level are excluded from the ground survey. Because of these issues, it should be emphasised that the changing nature of street-services should not necessarily be expected to parallel the wider and changing roles of services in the economy.
The discourse surrounding the changing street, in particular that of the Retail Revolution, has focused on the growth of multiples and the concomitant restructuring of the retail street. To study the penetration of chains, franchises and associated forms of capital over the time period of the study, the occupants of every site were initially manually classified into chains where appropriate.
To complement the identification of chains, it was decided to explicitly identify every situation where a named occupier was present in two or more sites anywhere within the two study Regions during the same year, by defining these sites as multiples. As such, when the terms chains and multiples are used here they are not meant to be read as synonymous, but as indicating fundamentally different classifications. It was decided not to define multiples on the basis of identically named sites within the same study Region, as it was felt that this was too restrictive a base for comparison, due to the relatively small sample size of both study Regions, and more importantly that comparing data within the study Regions wilfully ignored the fact that the two study Regions are hundreds of meters apart from each other. It must be noted that the identification and classification of these multiples is directly linked to the overall size of the study site: if the study had been based upon the data from four Goad maps, for example, it is expected that the overall numbers of multiples would be somewhat higher.
All site uses were thus divided into three broad categories:
| Chains/ Multiples/Franchises | Other | |||
| Boots | Greg’s | Banks etc. | All institutional uses | |
| McDonalds’ | Waterstones | BT | Vacant sites etc. | |
| Principles | RS McColl | Scottish Gas | Individual/sole traders | |
| Vantage Pharmacy | American Express | |||
Table 4.11: Sample capital classifications
From this dataset it is possible to study both the penetration of organised capital into the street landscape, and to develop a greater knowledge of the evolving street of chains, including how quickly they penetrate the streetscape, how geographically concentrated they are, whether they still represent an increasing percentage of street uses, and whether new chain sites are now replacing existing chain sites or non-chain sites.
There are many different ways of both defining and in turn measuring homogeneity or similar relative measures of sameness. While it is not possible to analyse how similar the streets within the study Regions are with other streets outside the study Region, it is possible to classify the data we already have to measure homogeneity in terms of the growth in numbers of chains and franchises in the landscape, as this represents a particular type of homogeneity, where basing analyses upon the numbers of chain sites provides a means of measuring the degree of similarity between the study Regions and the wider economic sphere. Using the particular definition of multiples outlined above provides an alternative analysis of levels of homogeneity within the study Regions, i.e. of all Occupants who are represented in multiple sites within the two study Regions in any given year. In addition, this data can be analysed to show us how varied streets are, and the numbers of unique occupants in each Area .
Just as we can measure homogeneity in terms of similarities in occupants, we can also measure it in terms of similarities in uses. Using the usage classification that was developed earlier (see section 4.5.1: Streets of Change), it is possible to develop indices of how varied the uses are for each street Area, and whether the variety of uses is changing and/or homogenising.
There is to some degree an inherent contradiction between the desire to impose a logic upon the analysis of the changing street, and what might best be characterised as the street’s unruliness and resistance to straightforward and simple explanations. As a result, any presentation of the results of analysis is by definition somewhat arbitrary, as there are always cross-references, points of difference, points of comparison, related and relevant ideas that cannot easily be shoe-horned into all the possible and relevant positions within a contiguous stream of data and commentary without creating logical holes elsewhere in the argument. It is a shibboleth that everything is connected/related, and this is true in this case as well.
The result of this has been a decision to separate (somewhat arbitrarily) the results into two chapters: Chapter 6, which will look at the broader question of change (or the lack thereof) qua change, while Chapter 7 focuses on more specific thematic aspects of the changing Edinburgh street.
Chapter 6 is divided into three broad thematic sections, focusing on the changing street, the stable street, and streets of homogeneity. The changing street is a broad overview of change, asking if streets are changing, how are they changing, are specific Areas changing faster than others, questioning whether vacancies can be used as indicators of change, and attempting to determine if there are identical structural patterns of change which may be used to indicate how the street and its role has changed during the study period.
In contrast, the section on streets and stability questions these discourses of change, and essentially reverses the questions asked in the section on change, by asking not how is the street changing but by identifying sitess of stability and focusing on the lack of change in the street. In particular, it focuses on the 1994 street, which as the final year in the survey provides the most appropriate opportunity to construct a ‘historic’ profile of the ‘current’ street. In addition, specific interest is shown in the sites from 1978 that have existed unchanged throughout the length of the study period.
Streets of homogeneity explores the questions of homogeneity and change that characterise many of the more abstract conceptualisations of the changing city. It uses the data on chains and multiples as indicators of homogeneity within and across the two study Regions, while the data on chains provides an opportunity to develop indicators of similarity and ‘sameness’ with the wider urban experience. In addition, data on Regional site uses is analysed to determine if there is evidence to support arguments about the increasing homogenisation of the street through decreases in the availability of a variety of services and goods.
To complement the broad analyses of change in Chapter 6, Chapter 7 focuses on the analysis of change in three specific themes: leisure, services and capital. Streets of leisure focuses on identifying the changing role of leisure in the study Regions, and determining how leisure is changing, if leisure provision is actually expanding, questioning whether there is a geographic concentration of leisure services. It analyses whether there is a general increase in leisure provision, or if specific types of leisure services are becoming more prevalent. The changing relationship between retailing and leisure is studied, as is the changing nature of food and alcohol consumption.
Streets of services focuses on the changing role of service provision in the urban street, and studies whether service sites are becoming increasingly prevalent. In addition, it specifically identifies institutional sites, and studies their changing role and impact.
Streets of capital focuses on the analysis of the changing roles of chains and multiples in the study Regions. After looking in detail at the changes of each of these classes of sites, it looks at the changes in the relationship between the two by studying the changing nature of both the chain-multiple and the non-chain multiple, before moving on to focus in detail on the changing retail scene and its relationship to the changing roles of chains and multiples. This analysis is then complemented by studying the links between leisure and service sites and chains and multiples.
