Place-name choropleths

From International Robin Hood Bibliography
Loading map...
Count of place-names per county. The darker green, the more the count for a given county exceeds the average for the 39 historic counties. The darker red, the further below average is the count. White represents values close to the mean.
Loading map...
This map represents the same data as that above but takes the widely differing areas of counties into account.
Loading map...
This map represents the same data as those above but attempts to take both areas and population densities of counties into account.

Template:PnNavMapEnglishPlaceNames

By Henrik Thiil Nielsen, 2017-11-06. Revised by Henrik Thiil Nielsen, 2017-11-19.

This page includes choropleth maps of the English counties, detailing the geographical distribution of Robin Hood-related place-names and localities from three different perspectives. The dataset on which they are based is set out in a table below.

The choropleth maps

A choropleth is something much more widely known than its name: a "thematic map in which areas are shaded or patterned in proportion to the measurement of the statistical variable being displayed on the map".[1] It thus resembles a heat map, but unlike a heat map in which variation is in principle continuous, it displays data averaged over discrete geographical regions, in this case the English historic counties. On all three maps, a bi-polar colour progression is used, counties with values above the mean – set to index 100 – being coloured a progressively darker green, while those below are coloured a progressively darker red. Areas with values close to the mean are white or nearly so. Ideally colour gradation should be linear, but since there are a few important outliers in the data, this could only be achieved with colour steps so small that they become difficult to discern or, on the other hand, such large intervals of values grouped together that resolution for values closer to the mean would be insufficient. Intervals are therefore larger for areas well above the mean. I give a brief description of each map below. Population and area data for England and the historic counties are taken from the 1891 census.[2] Many place-names first figure in mid-19th century MS tithe awards or printed O.S. maps, so it would be natural to pick an earlier census as data source, but that of 1891 is the first one that provides a complete (surviving) dataset.

Simple county count

For the first map, the count of English place-names and localities currently covered on IRHB is set to index 100 and used as the base for calculating the indexes for individual counties. Counties with an index value between 100 and 109 (both inclusive) are shown in white, those with higher values in progressively darker shades of green, those below 100 in increasingly dark shades of red, Since the counties vary so widely in size, we are here, as it were, comparing gooseberries to water melons as well as fruits of all sorts of intermediate sizes. Yet the main tendencies recur on the next two, more meaningful representations of the data.

Taking area into account

For the second map, the density of Robin Hood-related place-names and localities in England – expressible as area in km2 per place-name or vice versa – is taken as index 100 and indexes calculated for individual counties based on their land areas and counts of relevant localities/place-names. Colour gradation as on the first map.

Taking area and population into account

The third choropleth attempts to take area as well as population density into account. Here we first calculate an expected place-name count for each county based on its area and the average count per km2 for England. On the basis of the population densities of England and the county in question we then adjust the expected figure upwards or downwards to reflect the population density of the county. This is done in linear fashion. The result is used as index 100 for the county, whose index value is then calculated on the basis of the actual count of place-names/localities. The colouring thus reflects the extent to which expectation is exceeded (or the opposite). I do not know whether a statistician would find this procedure reasonable, but it obviously takes people to give places names, and it seems a reasonable assumption that population density has at least some influence on the density of named localities in a given area. I doubt if this effect is linear, but as I have no relevant empirical knowledge, any adjustment I could make to the way in which it is factored in would be arbitrary. Colour gradation as on the preceding maps.

Notes and qualifications

While figures in the dataset table are updated once a day, it may take two or three days before the choropleths are updated. So when new material has just been added to the place-names section of IRHB and this ought to cause changes to the choropleths, you may find the table and maps slightly out of sync for a day or so. Keep in mind that the choropleths reflect the current state of my work. An estimated 75 to 150 place-names are still to be entered, many of these, mostly of fairly recent origin, belong in the counties that are currently dark red on the maps.

Dataset and colour codes

Click header to expand dataset.

<xmldoc path="bot/includes/kml" file="counties.xml" stylesheet="table.xsl"></xmldoc> <xmldoc path="bot/includes/kml" file="counties.xml" stylesheet="colors.xsl"></xmldoc>

Background

Notes