GEODEMOGRAPHICS | OAC | MOSAIC UK | OTHER PROPRIETARY CLASSIFICATION SYSTEMS
One of the most succinct definitions [1] of geodemographics captures effectively the geographical nature of neighbourhood classifications (Sleight 1997, p.16): "geodemographics is the analysis of people by where they live". These few words set phenomenon in space, and imply a cause-effect relationship between places and people. This link is bi-directional, in a sense that places with the same characteristics and behaviours attract people of the same type and that these characteristics, in their turn, reinforce neighbourhood characteristics (Harris et al. 2005). In a classification framework populations that fall into the same category are likely to live in similar neighbourhoods that may nevertheless be located far from each other [2].
Statistical techniques help to segment people sharing similar attributes in clusters. There are different methods available for clustering (Everitt et al. 2001) but one of the most commonly used is the k-means algorithm, which successively allocates and reallocates observations until a desired statistical outcome is achieved - that is, a classification of N zones in k number of clusters. The N zones may be area units [3] at the following fixed geographical levels:
- Administrative units: wards, counties or local authorities
- Statistical units used for disseminating census data: Ouput Areas (OA), Lower Layer Super Output Areas (LSOA), Middle Layer Super Output Areas (MSOA) or Upper Layer Super Output Areas (USOA)
- Postal units: postcode sector or unit postcode
Most of the data used in geodemographic cluster analysis are derived from National Statistics [4], and profiling areas according to population demography plays a key role alongside economic (e.g. income, occupation) and social (e.g. wealth, education) measures. As a general rule, the greater the number of variables available, so the richer the classification will be. The choice of the number and the type of variables to be classified, and the weights to be assigned to them, rests with the analyst.
The assignment of weights is particularly important, and is argued by some to be the most skilled step in the classification process: to the detractors of geodemographic classifications it also gives rise to the most trenchant criticisms, since the weighting procedures of commercial systems are usually hidden from public view. Weights are a sort of "magical ingredient" that makes each neighbourhood classification different from each other even when using the same data.
[TOP]
OAC - OUTPUT AREA CLASSIFICATION
The Office for National Statistics (ONS) distributes a national area classification system based on Census 2001 variables known as the OAC (Table). It uses 7 Supergroups, which can be subdivided into Groups and Subgroups (Table). Each level is a result of K-Means clustering of the remaining data structure (Vickers et al. 2005).
Most of Southwark area falls into Supergroup 7: Multicultural. This Supergroup divides into two Groups: Asian Communities (7a) and Afro-Caribbean Communities (7b), where the latter is the dominant ONS area class in Southwark.
Census variables used in ONS area classification system [TOP]
| Age 0 - 4 | Rooms per household |
| Age 5 -14 | People per room |
| Age 25 - 44 | HE qualifications |
| Age 45 - 64 | Routine/Semi-Routine occupation |
| Age 65+ | 2+ Car household |
| Indian/Pakistani/Bangladeshi | Public transport to work |
| Black African, Black Caribbean or Black Other | Work from home |
| Born outside UK | Long Term Limiting Illness (Standardised Illness Ratio) |
| Population density | Provide unpaid care |
| Divorced | Students (full time) |
| Single person household (not pensioner) | Unemployed |
| Single pensioner household (pensioner) | Working part-time |
| Lone parent household | Economically inactive looking after family |
| Two adult no children | Agriculture/fishing employment |
| Households with non-dependent children | Mining/quarrying/construction employment |
| Rent (public) | Manufacturing employment |
| Rent (private) | Hotel & catering employment |
| Terraced Housing | Health/social work employment |
| Detached Housing | Financial intermediation employment |
| All Flats | Wholesale/retail employment |
| No central heating |
ONS Area Classification Supergroup and Group titles [TOP]
[TOP]
MOSAIC UK
Mosaic UK is a geodemographic segmentation system developed by Experian. The mainstay of the system is the 2001 Census, supplemented by an assemblage of additional data derived from the Electoral Roll, Experian's Lifestyle Survey, measures of Consumer Credit Activity, the Postcode Address File, various registers of shareholders, house price, council tax and Office for National Statistics (ONS) local statistics.
The classification system is constructed through cluster analysis of these variables and assigns the UK population into 11 Groups and 61 different lifestyle Types. Detailed descriptions or 'pen portraits' exist for each Type. Mosaic divides the UK population into the c. 1.7 million unit postcodes with an average of 15 households in each. Experian is in the process of coding Mosaic Types to various national surveys.
[TOP]
OTHER PROPRIETARY CLASSIFICATION SYSTEMS
There are a great number of different classification systems for geodemographic targeting available on the market. For an excellent review of the main systems available refer to Peter Sleight's book Targeting customers. How to use geodemographics and lifestyle data in your business (Sleight 2004). Below is a list of the most diffused in the UK with links to pages that contain information on each classification system:
REFERENCES [TOP]
Everitt, B. et al. 2001. Cluster Analysis. London: Hodder.
Harris, R. 2003. Population mapping by geodemographics and digital imagery. Remotely Sensed Cities: 225-241.
Harris, R. et al. 2005. Geodemographics, GIS and neighbourhood targeting. Chichester: John Wiley & Sons Ltd.
Sleight, P. 2004. Targeting customers. How to use geodemographics and lifestyle data in your business. Henley-on-Thames: WARC.
Vickers, D. et al. 2005. Creating the national classification of census output areas: data, methods and results. Leeds: University of Leeds, School of Geography.
Notes [TOP]
[1] Richard Harris (Harris 2003, p.225) approached a possible definition in a more formal and scientific way stating geodemographics as "the analysis of socio-economic and behavioural data about people, to investigate the geographical patterns that structure and are structured by the forms and functions of settlements" .
[2] From a statistical point of view these assumptions made geodemographics a powerful tool to explore but not to perform tests and validate hypotheses.
[3] Clustering techniques can also be applied to individual data, but geodemographics are typically based and defined as small area classifications.
[4] Most commercial suppliers often integrate additional information coming from either public or private sector sources.
