The term ‘data visualisation’ is often associated with unusual or complex (and often artistic) visual representations of data. It’s easy to be drawn to these images, and to overlook the value of more traditional and familiar techniques which offer tried-and-tested ways to spot trends and patterns in your data. If your aim is insight rather than art, simple and familiar visuals work best:
1. Comparison. A bar graph is a familiar means of making numerical comparisons between variables. Whether you wish to see a side-by-side comparison of journal income, customer interests, or email alert topics, a bar chart provides the easiest and most accessible way to compare the relative size of each value. Making each bar ‘clickable’ to search for all customers associated with each bar makes it really easy to move from insight to action.
2. Change. Line graphs are the ‘classic’ approach for visualising change over time and for trend analysis. It can be useful to assess many publishing variables in this way, e.g. author submissions, alert signups, online registrations, number of print and online subscriptions etc. Again, making the chart ‘clickable’ to zoom in from a yearly view to a monthly view, or to link directly to all related customers, helps to bring the chart to life.
3. Proportion. Pie charts are a great way to visualise the division of a set of customers or contacts into categories, and to then ‘dig’ into sub-groups of interest. This might be achieved using category data from your own systems (such as registered fields of interest, job roles, subscription types, and rate codes) or using profiling data from third-party resources (such as Ringgold Identify’s ‘sector’ and ‘type’ fields).
4. Relationship. Sometimes your interest lies in seeing two variables plotted together, e.g. author submissions by country. A cross-tab with visual highlighting of clusters in the data can be a good way to present this, graded ‘hot spots’ helping to identify where best to focus your marketing activities. In addition, there might be other ‘relationships’ that you’d like to see and explore, such as those between a consortium, its member institutions, and affiliated individuals. Interactive family tree-type visualisations can be extremely useful here.
5. Overlap. Rather than run a complicated search for a specific segment of interest, you might prefer to define a number of smaller, focused searches, and then combine the resulting sets to visually explore the overlaps and non-overlaps between them. Venn diagrams are an excellent visual means to explore the intersections between different customer groups, for example, alerts, pay-per-view, and subscriber data. Access to any overlap or non-overlap is then just a click away.
6. Location. Geography is of course a significant variable, especially for publishers with regional sales teams. Location maps can be colour-coded at the country or state level to reflect data values, helping to draw out any spatial patterns in your data. This can be useful when sales teams are planning regional trips and site visits.
The DataSalon Blog 