Libraries Hacked

Promoting open data and digital prototypes in libraries

9 minute read

What is a library catchment area? Is it the area where people are reasonably able to get to the library? For example, a certain mile radius or travel time? Is it based on home locations of members of the library, or those who have been recent ‘borrowers’? Or, is it the area from which ALL visitors to the library may come?

For library services, defining a catchment may depend on what they wish to achieve.

  • to assess library locations (static and mobile) to see if these are spread appropriately;
  • to understand which demographics (e.g. based on income or age) are most likely to be users of a library;
  • to understand which communities aren’t currently library users;
  • to tailor library services based on existing membership;
  • to tailor library services based on geographic location.

Analysis of a service is often done prior to a consultation exercise. But geographic intelligence, which in turn gives intelligence about people and services, is always important. It is necessary to understand the people a library serves (or should serve), and the different requirements that may go along with that.

Introducing geography to ALL library-based data (e.g. simple monthly issue/visit stats) could stop these being ‘one-off’ exercises. The results of catchment profiling are often used for many years, rather than being treated as an ever-shifting profile.

What have library services done in the past to define catchment? This post will look at 3 examples.

  • Gloucestershire. Library provision 2012.
  • Bedford. Library profiles 2014.
  • Sutton. Consultation 2014.

Before that, here’s some jargon.

LSOAs</strong>. Lower Super Output areas. These are small areas of an authority (smaller than a Ward or Parish), created from census data in order to profile that area. Such as by deprivation, or likelihood of owning a car. They tend to have roughly the same number of people (around 1500), so are good to use in comparisons. This is unlike a Ward or Parish (or even postcode), which tend to be more randomly populated.

OAs. Output areas. These are base unit census areas. They are smaller in size than LSOAs, but used as the ‘building blocks’ that make up the LSOAs. They consist of around 300 people.

Indices of deprivation. These are statistics on deprivation, released by the ONS, for each LSOA. There are around 32,000 LSOAs in England, and each is ranked on a number of deprivation measures (e.g. income, or education). A ‘multiple’ deprivation measure is also calculated which is a weighted combination of all the deprivation measures. For example, a community east of Jaywick near Clacton-on-Sea is defined as the most deprived in all of England based upon this index of multiple deprivation.

Mosaic. This is a commercial demographic profiling tool from Experian, the credit rating people. It profiles households into 15 main socio-economic groups and, within this, 69 types. The groups are things like ‘Happy Families’ (families from Middle England). A type may be more detailed like ‘Upwardly mobile South Asian families living in inter war suburbs’. The data will also give indications such as whether a household is likely to take advantage of free credit, which is why the tool is used frequently by financial marketing departments. Other products are available.

Gloucestershire 2012

In 2012, Gloucestershire developed a library strategy, involving catchment areas. But Gloucestershire is big. Before creating a catchment area for each library, they created 6 main districts. Each of these would have a set of libraries, then assigned their own catchment. The 6 districts are Gloucester, Cheltenham, Stroud, Tewkesbury, Forest of Dean, and the Cotswolds.

What methodology was used to create the library catchment areas?

the Library Service wanted to understand the characteristics of the communities served by each library. In consultation with library managers, catchments were drawn up using COAs as a basic unit. The catchments reflect natural communities and local issues such as public transport, main road links and shopping facilities rather than being strictly district based. Using these enabled the Research Team to report Census 2001 information for each library.

This appears to be a partially manual exercise, using expertise from staff and other intelligence to decide which output areas (COAs = Census Output Areas) are assigned to each library. This could be quite time-consuming, as Gloucestershire consists of hundreds of Output Areas. But it’s likely the majority would be fairly obvious (the immediate area surrounding a library). It’s interesting to see this being done using the expertise of those running libraries. Often the task can be effectively outsourced to GIS departments, or research intelligence teams. But it is staff who will know the most about their own users and local area.

In the documentation there is considerable profiling of each catchment area using stats such as indices of deprivation, access to vehicles, protected characteristics, and more.

To summarise, Gloucestershire:

  • created catchment areas with no overlap between different libraries (each community is assigned a single library);
  • used expert knowledge to assign areas;
  • did not base catchment areas on membership or usage (aside perhaps as an aspect of expert knowledge).
  • ensured each area of the authority was assigned to a library.

Gloucestershire County Council Library Service: Context and User Needs Assessment

Gloucestershire Library Strategy 2012

Bedford Library Profiles 2014

In 2014 the Bedford Borough Council Community Intelligence team profiled Bedford’s libraries. This included defining, for each library, an ‘all users main catchment area’, and an ‘IT only catchment area’. Those areas were then profiled using the indices of deprivation and Mosaic.

Although these are detailed profiles, they don’t include how the catchments were defined. On querying this, Bedford said the individual who did the work has left, and that knowledge has left with them. These things happen! It is clear though that the two catchment areas are based on usage data, and taken from member address information. That membership has been analysed to define IT users, and All Users, in order to construct two different geographical areas.

(This is also a good example of where, if such work had been openly documented and shared, it would be repeatable and useful for others, internal and external).

This is the opposite to the method used by Gloucestershire, who take the geography of the whole authority to split up areas, regardless of current usage or membership. The Bedford profiles are worth studying for the disparity between the different types of catchment area (IT Only/All users).

To summarise, Bedford:

Sutton consultation 2014

Sutton created profiles of each of their libraries, defining catchment as:

all Lower Super Output Areas (LSOAs) where 5% or more of the population borrow from the library.

A borrower is normally a library member who has loaned an item from the library, often filtered within a certain timeframe e.g. a year.

There are disadvantages to this method. It takes no account of users who use the computers, WiFi, or other materials. It doesn’t guarantee that all areas of an authority will fall into the catchment of a library. For providing a needs analysis across the authority, this can be questionable. Regardless of current membership, the authority provides a service for all residents. It is reasonable to assume everyone should fall into the catchment of at least one library.

It does allow for catchment areas that overlap between libraries. In this way, Sutton Central library catchment covers 91 of the 121 LSOAs in Sutton, showing that a Central library will likely be used by the whole authority. This seems in some ways more realistic than an approach which neatly divides an authority into distinct areas.

A problem occurs when a library does not have any LSOAs where 5% of the population are borrowers of that library, and therefore has no catchment. This was the case for Beddington library. Sutton decided:

The catchment area for Beddington Library is therefore taken to be the two LSOAs which contain the highest numbers of active Library members.

That seems an arbitrary and problematic definition. For one thing it pre-defines the population count in that catchment, as 2 LSOAs will be around 3000 people. This figure is then used later on in the analysis, in a statement that says the population of Beddington’s catchment area only amounts to 1% of the population of Sutton.

Beddington library was actually situated near the intersection of 3 LSOAs. Because the decision only allowed for 2 to be included, this meant the third community was excluded from that library catchment. Looking at this on a map it leads to a slightly absurd situation where a location across the street from the library was not considered part of its catchment.

Sutton also do some driving distance analysis. A ‘10 minute drive area’ (2 mile radius) is plotted around Beddington library to suggest alternative libraries that could be reached by car, of which there are 3. The suggestion is Beddington users could drive elsewhere in 10 minutes. It would be more appropriate to plot the drive radius from the centre of the catchment area, to represent the distance required to travel by the population. The majority of people would have to travel further than described.

In September 2015 a proposal to Sutton residents was put on closing Beddington library.

Beddington library’s catchment area does not score highly on measures such as the indices of deprivation and the needs analysis. In particular, the area is well served by public transport and 84% of people in the area have access to a car/van.

Beddington library was closed in 2016. In December 2016, Sutton were featured as a case study in the Libraries Taskforce Ambition document.

The consultation on changes to Sutton libraries was backed by research mapping demographic trends and needs analysis across the borough. This was so staff had a clear view of the communities the libraries serve.

To summarise, Sutton:

  • Defined catchment based on ‘borrrower’ members
  • Created overlapping catchment areas, showing how the Central library is used by residents across the authority.
  • Used a definition of catchment area that did not work for all libraries - forcing inconsistent definitions within the same study.

  • Used a methodology which could exclude various communities from the analysis.

  • Libraries and culture downloads
  • Beddington Library - Facts and Figures

Conclusions

There is clearly no right or wrong way to define library catchments. But from looking at different examples it’s possible to make some suggestions.

  • If defining catchment areas based upon existing usage and membership it’s important to use more than just borrower data. Bedford showed that including IT Only users created very different catchment areas.
  • A definition of a catchment area should be consistent. The Sutton definition of a catchment area didn’t work for all libraries.
  • If assessing the needs across the authority, it is important that ALL areas of the authority are included in catchments.