Postcode Search: ArchSearch and geoXwalk

In October we launched a new 'Postcode' search within ArchSearch, based on tools provided by Edinburgh University. James Reid, project manager, describes this work and why it is more sophisticated than the simple search box suggests.

Postcodes are one of the most familiar types of geography by which users can initiate spatial searches and provide a simple geographical access point to locate resources related to relatively small areas. They provide a powerful filtering tool and enable searches such as 'what significant archaeology is there close to my home address?'

The postcode search in ArchSearch is powered by a facility at EDINA (http://edina.ac.uk) known as geoXwalk (www.geoxwalk.ac.uk - pronounced geo-cross-walk). geoXwalk enables geographic searching of resources within the UK academic digital library by providing critical middleware infrastructure through which diverse digital libraries may undertake geographic querying.

It has long been recognised that space provides a powerful mechanism by which to search. There has been some recognition of this within digital libraries with efforts to provide geographical indexes to online resources. However, such efforts have generally been piecemeal and suffer from being too restrictive in the concept of 'geography' they utilise. Placenames are frequently used for simple text based searching but these tend to be 'blunderbuss' type searches. Consequently result sets are poorly discriminated and often return spurious results. This is a simplistic, 'lowest common denominator', approach to geographic searching for a number of reasons.

ArchSearch postcode searches powered by geoXwalk.

ArchSearch postcode searches powered by geoXwalk.

Firstly, there is little consistency in the geographic vocabulary used. In the UK, resources might be indexed (if at all) by placename, postcode, administrative code or other arbitrary coding convention. This has led to a plethora of geographies. In 1990 a sample audit of resources by UK Data Archive estimated that over 90 different coding conventions and practices were in use. Secondly, coding schemes are also subject to revision and the geographical entities they describe change spatial form on different temporal frequencies. UK electoral wards change frequently and often quite dramatically whereas county boundaries are relatively consistent over longer periods. In other words, UK geography is complex!

In practical terms what this means is that geography as a search dimension lacks both consistency and persistence which makes reliability, repeatability and utility in searches problematic. There is however one geography which obviates all these problems, namely the use of a coordinate system (such as the UK Ordnance Surveys' National Grids or, alternatively, a latitude/longitude encoding). This is already widely used in archaeology and supports the map search facility in ArchSearch. By encoding coordinate representations of our various geographies we thus arrive at a geography that is both consistent and persistent and provides a baseline against which comparisons of various geographies may be conducted. This is the key to geographically enabling resources and is what underpins the ability of geoXwalk to provide answers to complex queries. As the name suggests, geoXwalk, can 'crosswalk' the varying manifestations of geographies to 'translate' one representation into another and, in the case of ArchSearch, this is done by a simple 'crosswalk' of a postcode to a coordinate 'footprint'.

James Reid
EDINA GeoData Services

ArchSearch
http://ads.ahds.ac.uk/catalogue/search/basic.cfm

In this issue ...