Best practice on using thematic maps
Main Content
What is a thematic map?
There are lots of different types of thematic map. However, on the Neighbourhood Statistics website, we use the term "thematic maps" to mean area-shaded maps. This type of map (usually referred to in cartographic literature as a "choropleth" map) portrays geographical areas (such as Regions or Super Output Areas) in a range of different shades, where each shade represents a different, banded value. The purpose of these maps is to present the data in a way that allows geographical patterns to be identified. For example, if you want to know whether there are particular Local Authorities with a high concentration of retired people you could map the proportion of the population that is over retirement age in each Local Authority.
The total area covered by the map will depend on the question you are trying to answer by drawing the map. For example, if you were interested in the pattern of income deprivation in a particular local authority you could map the income deprivation score from the 2004 Index of Multiple Deprivation for all Lower Layer Super Output Areas within that local authority. We give an example of this below for the city of Birmingham.
What sorts of data are best for a thematic map?
One of the key purposes of thematic maps is to demonstrate how data vary geographically.
This means that, in most cases, you should use data that has been standardised in some way so that direct comparisons between areas can be made.
What do we mean by 'standardised'? Consider the example of the retired population again. If you are interested only in whether one area has a higher number of retired people than another, then mapping the number of retired people in each area is sufficient. However, the differences that you see between areas may arise simply because there are more people living in some areas than there are in others.
In general, if you want to explore geographical patterns it is better to take account of the total population in the values that you map. To continue with the example of retired people, if you divide the number of retired people in each area by the total population in that area (by doing this you are standardising on the total population) and map the results, this will allow you to see where there are high concentrations of retired people in relation to the total population. This is a better measure of the relative importance of particular areas. Dividing by the total number of people has standardised the data so that the comparison takes into account variations in the number of retired people that might be expected as a result of variations in the total number of people.
On the Neighbourhood Statistics website data that have already been standardised will mostly have a 'measurement unit' of:
- Percentage - e.g. dividing the number of over 65's by the total number of people in the area and multiplying by 100 will give the percentage of the population who are over 65
- Rate - e.g. dividing the number of offences recorded by the number of people in the area and multiplying by 1000 will give an offence rate per 1000 population
- Average - e.g. dividing the number of people in an area by the number of households will give the average number of people in a household
- Score - e.g. Index of Multiple Deprivation 2004
- Density - e.g. number of people divided by size of area gives a population density
Are there any exceptions?
Super Output Areas (SOAs) are more consistent in population size than administrative geographies (such as wards, local authority districts). If counts are mapped by SOAs, and you are interested in the proportion of the total population who are over retirement age, mapping either the number or percentage of people who are over the retirement age in the area should give similar results. This would not be the case if the boundaries used were wards. The Lower Layer SOA is the most consistent in population size and is therefore the most suited to this approach. The Middle Layer SOAs have a more variable population size.
