In 2004, Manchester Geomatics developed a web-based intelligence system,
containing a comprehensive set of spatially referenced information for the Manchester Salford Pathfinder (MSP).
This system was used in the RFI exercise to support local level monitoring and
evaluation, and in developing its funding bid.
The aim of the project is to revitalise underperforming areas of Manchester and Salford through a process of investment and
regeneration.
The MSP research team recognised that long-term evaluation of the performance of the project would be difficult with existing
geographies as the underlying distribution of housing and industy would be fundamentally changing during the regeneration process.
To help with the evaluation, Manchester Geomatics developed a suite of geographies that are stable over time, but still respond to
underlying changes in the locations of housing or industry. Many key indicators have now been mapped to these geographies.
The image below summarises the two level geography: block bits and block neighbourhoods, shown above.
So what is wrong with existing geographies? Existing geographies are either too static (such as Census geographies -
fixed in time since 2001) or too dynamic (Postcodes change to better accomodate the distribution of mail by Royal Mail,
even though the underlying distribution of dwellings and places of work have not changed!). Some estimates show that 15% of
postcodes have changed nationally since 2001, and that the vast majority of these are "churn" - one postcode replacing another.
Also, size and shape are issues. In terms of "shape", Postcode based geographies (including the census geographies) do not map
cleanly with real world boundaries.
Sizewise, MSP decided that a geographic level of approximately 50 households per area to better reflect small local changes.
Postcodes were deemed too small (on average 15 households) and would break confidentiality restrictions and Output Areas
(on average 125 households) were thought to be to large and might mask early indications of change.
Compare the following examples of each geography:
Street (Chorlton)
| |
Street (Harpurhey)
|
|
Street (Harpurhey)
|
Postcode (Chorlton)
| |
Postcode (Harpurhey)
|
|
Postcode (Harpurhey)
|
Output Area (Chorlton)
| |
Output Area (Harpurhey)
|
|
Output Area (Harpurhey)
|
Block (Chorlton)
| |
Block (Harpurhey)
| |
Output Area (Harpurhey)
|
When developing the geography, we realised that a single "do everything" geography was impossible. What was necessary was
a two level geography. The lower level, block bits are much more suitable for data collection and are logical units that relate well
to addresses, streets and postcodes. The upper level, block neighbourhoods, are collections of block bits with similar characteristics,
that meet minimum numbers of dwellings or organisations to adhere to confidentiality guidelines. As such they are ideal for presentation
as well as analysis of the information.
Here are some examples of Block Bits and Neighbourhoods for the same areas shown above:
Block Bit (Chorlton)
| |
Block Bit (Harpurhey)
|
|
Block Bit (Harpurhey)
|
Block Bit
| |
Block Bit (Harpurhey)
|
|
Block Bit (Harpurhey)
|
Block Neighbourhood
| |
Block Neighbourhood
(Harpurhey)
|
|
|
Overall, the creation of the geographies was a success. MSP made use of them for collating, evaluating and
presenting local information.
The block bits and block neighbourhoods are starting to become adopted
more widely. The Office for National Statistics (ONS) are currently evaluating their suitability for use in the 2011 census.