Inputs to LGAS
If you upload a postcode or 1991 census ED then LGAS will add cluster codes to your file and show the
frequency of your data by cluster. If you input a file of postcodes or census ED codes and a single
attribute count or score variable then LGAS will also show average values per cluster. If you input a
file of postcodes or census ED codes and two values (a numerator and a denominator) then LGAS will report
rates per cluster or create a gains chart. There are also other
analysis options you can try out, as contained in the help file.
Input Data Formats
A. 1991 Census Enumeration District Codes
You can input data using either 6 or 8 character 1991 census ED codes. The full 8 character code consists of:
whilst the 6 digit code exploits the fact that the county code is redundant. Note the hierarchical nature of
the code, in that ED is nested within wards which are nested within districts which nest within counties.
In fact the structure of the census ED code is one of the neatest aspects of the census.
- 1-2 County code
- 3-4 District code
- 5-6 Ward code
- 7-8 ED number
B. Unit postcodes
You can input data for full postcodes. However there are a number of complications:
- The 1991 census EDs for England and Wales do not follow postcode boundaries;
- Postcodes change; there are new ones (births), deaths, and recycled dead ones. As a result the
postcode - ED look-up tables change. So not all postcodes that existed in 1991 will still exist and
some will have been changed due to re-organisation. Additionally new ones will have appeared.
- There are three different versions of the postcode to ED tables for England and Wales.
- A single postcode can belong to more than one census ED in England and Wales. The record is 10! The
problem is that there are different versions of the ED to postcode directory for England and Wales making
it very difficult to know how to share out the overlapping ED codes. The simplest solution is to ignore
the problem and assume away any errors. After all the postcode to ED lookup process will create about
10% mismatches anyway so what's a little bit more error going to do? Probably not much! The effects are
probably going to be extremely small.
C. Atttribute Data (Numerator and Denominator)
You can use LGAS to segment any data values of interest. Examples would be:
D. Sample Formats
- Deprivation Index: is it concentrated in certain residential area types?
- Burglary count and Total Households
- Disease count and Population at Risk
- Mailing Response count and Population Mailed
- Ownership of something with population at risk of owning
You can upload either comma, space or tab separated data with or without a single line header. For example
01AAFQ07 or AAFQ07 for an ED code
AA7Q07,56.6 for an ED with a single data item
AA7Q07,55.6,123.7 for an ED with a numerator and denominator
AA7Q07 55.6 123.7 same as above but space separated
Return to: Help Topics LGAS homepage