Examples of graphics developed by giCentre and used by Leicestershire County Council
Click the thumbnail graphic or the file name to download a full size image
Please contact Jason Dykes with queries, for additional explanation or to discuss further.
Visualization at giCentre
Researchers at City University London are developing innovative graphics to show patterns and relationships in large amounts of data. The graphics do so in engaging and useful formats that can help inform citizens and decision makers and add value to a range of data sets.
Interactions are used to query the information, provide details on demand and make comparisons.
Prototypes are being developed to enable citizens to act on government information, compare authorities and hold their local decision-makers to account through the Timely Information for Citizens. The initial pilot is being developed to provide citizens with access to local and national results of the Place Survey through engaging graphics.
This work builds upon work undertaken by Leicestershire County Council and the giCentre at City to use interactive graphics in research, reporting and decision-making and evaluate their effect. The pilots are in development and will draw upon techniques developed by the giCentre Some of these are shown and explained below using statci graphics.
1. Spatial TreeMaps: Egalitarian Mapping of People in Places - IoD Example
Where large numbers of people live in small areas, as is typically the case in cities, traditional maps graphically under represent these people - they don't give all people the equivalent space on the page that they deserve. Our spatial treemaps allocate the same amount of graphical space to all citizens whilst retaining the hierarchical geography of the units they live in.
We have used them with Leicestershire County Council to clearly map the 2004 and 2007 Indices of Deprivation at Super Output Area level. Here, output areas nest within SOAs, named wards and the seven districts - all sized according to population.
We can use colour to show levels of deprivation or change within this novel map. Here 'Crime' is mapped and the blue shades show a decrease in ranking across NorthWest Leicestershire that is not evident elsewhere.
Our interactive methods allow analysts and citizens to vary the index that is displayed and identify particular areas and see their rank and how this has changed.
An SOA in Sileby ward is selected and its change in rank from 21,000 to 6,000 highlighted on the associated 'rank ladder', which shows change in rank for the Crime index between 2004 and 2007 for all Leicestershire SOAs.
This kind of interactive graphic enables citizens to compare their OA, SOA, ward or district with others according to the various indices and see how things are changing within these localities. It enables decision-makers to evaluate the impact of policy and identify areas at different resolutions that may benefit from action.
2. ODMaps: Effective Mapping of Flows between Origins and Destinations - Travel to Work Example
When examining flows between an origin and a destination, for example migration, livestock movements or commuting patterns, a simple mapping solution is to plot lines between pairs of origins and destinations. clg-ttwFlows.png to the left shows this for travel-to-work patterns for the Leicestershire region. This form of mapping has two problems. Firstly it is not possible to identify the direction of flows (are more people travelling from Leicester to Loughborough or Loughborough to Leicester?). Secondly longer flow lines obscure the shorter lines beneath them (how much commuting occurs within Leicester?).
The Origin-Destination map, or OD map overcomes these problems. The region of interest is divided into a grid (in these examples Leicestershire is divided into a grid of 12 by 12 cells). The volume of commuting originating from each grid cell is then calculated. Each grid cell is then subdivided into a further 12x12 cells representing a small map of Leicestershire. On each small map, cells are coloured according to the number of flows that originate from the large origin cell and end in any of the 144 destination cells (darker red = more commuters). Thus we have a collection of small maps of work place destinations arranged according to the home locations of commuters (see clg-ttwOrigins.png).
We can easily swap origins and destinations to give us a collection of maps of commuters - home locations arranged by their workplace location (see clg-ttwDestinations.png). This allows us to determine, for example, that c.1000 people travel from Leicester to Loughborough to work compared with c.1700 from Loughborough to Leicester.
3. DashBoards: Shaping Decision-Makers Views - Performance Example?
Dashboards are the visual interfaces to the most important information needed by decision-makers that enable it to be monitored at a glance. They provide an effective means of summarizing and comparing large amounts of data - such as performance data in this case.
Senior managers and politicians within Leicestershire County Council use dashboards to display performance data for each corporate objective so that possible performance risk can be easily identified for action. Politicians have responded positively to our dashboards, extolling their virtues at council meetings.
4. Quartile Distance Maps: Geographic Summaries of Point Distributions - How Far do a Leicestershire Library's Best Customers Travel?
Leicestershire Library Services have been interested in analysing their user records to identify geographic and other trends. We're working with LLS to develop interactive graphics that help them use these valuable data to understand library usage patterns and inform decisions.
Here we map the origins of the most recent and frequent customers. This is a busy graphic and it is difficult to detect trends due to the overlapping lines and densely populated origins.
We summarise the patterns with quartile distance maps. Here grey symbols show library locations. These are sized according to numbers of users. Red circles are then plotted at the same locations to show the distances travelled by the 25, 50 and 75% closest customers.
The maps allow us to see patterns amongst the 400,000 library users in Leicestershire. We can differentiate between large libraries with local usage patterns and small libraries with a more regional usage base. We could produce similar maps for other types of usage: least frequent users, particular social groups, particular types of borrowing activity, etc.
We can use a spatial treemap to show information about the origins of library users. Here the home locations of library users are mapped at output area level. Output areas are sized according to population and coloured according to whether the OA has more of fewer 'best users' than expected given the population. The redder the OAs the greater the number of best users, the bluer the fewer. This is a simple model that could be refined to account for travel distance, type of library usage or population characteristic, but it has already helped us identify areas where library usage is lower than might be expected.
Our interactive applications enable Library services to switch between representations, query them for details and map a range of data sets relating to libraries and output areas.
5. Spatial TreeMaps: Egalitarian Mapping of People in Britain - OAC Example
Spatial Treemaps allow us to map large number of geographic units concurrently and relate information about them.
Here we show all 1,526,404 postcode units in England, Wales and Scotland with residential delivery points concurrently, sizing each by residential delivery points (as a proxy for population) and arranging in a way that maintains geographical relationships and the hierarchy of postcode area / district / sector / unit. See how the Glasgow (G), Newcastle (NE) and Exeter (EX) postcode areas are arranged in the maps on the left.
OAC (Output Area Classification) classifies the UK population into categories that represent their socio-economic characteristics. There is a 3-level hierarchy of categories: 7 super-groups, 21 groups and 52 sub-groups. The maps use colour to show the most characteristic sub-group of the population by postcode unit.
The following legend relates colours to OAC groups and postcode areas to their geographic location. The size of each sub-group, group and super-group on the left shows how many places fall into each of these categories across Britain.
Units in the map are coloured according to the sub-groups of the Output Area Classifier, enabling us to see how the groupings (sub-groupings and super-groupings) most associated with areas at different levels of geography vary.
These kinds of carefuly arranged dense graphics enable us to see patterns in huge amounts of infrmation and compare and relate local, regional and national patterns. Here we are graphically assimilating detailed information about 60 million people recorded in 40 census variables in more than 1 million places.
We provide the maps in landscape and portrait format to fit to screen or page.
6. giCentre at City University London
The approaches developed in the giCentre are informed by empirical and theoretical work in the domains of Cartography, Geographic Information Science and Information Visualization.
The giCentre is involved in developing approaches to effectively gather visualization requirements and to evaluate the use of visualization in applied settings.
We have received recognition for our work from the national and international research communities.
Several of our interactive visualization applications are available for you to try online.
- HousePrices: explore temporal variation in house prices and sales across London by house type.
- Traffic Patterns: find out how traffic speeds vary in London by time, day and month in out interactive visualization of 42 million GPS points collected by eCourier vehicles.
- BookScraper: compare the language used by classic authors in our interactive visualizations of book content.
- RAEViewer: Engage with the results of the recent Research Assessment Exercise through an interactive graphical representation of the 67 Units of Assessment.
References to academic publications describing our work are also available through our research pages.