Thinking about MI in the form of multi-dimensional cubes provides a very useful means of classifying the various distinctions that describe the data we are working with in graphs and tables.
By way of example, let’s look at the dimensions of the data in the table below:
The table contains some summary financial results for the Total UK, which is made up of four regions (North, South, East and West). This geographic distinction represents the business units of our company and can be represented as one of the edges of a cube.
The rows of each part of the table represent the lines of the P&L account (or a KPI calculated from two of those lines). This dimension of reporting items represents another edge of our cube.
The first set of columns in our table represent different versions or ‘modes’ of data (actual, budget, prior year etc.), and we can use this to complete the dimensional structure of our cube.
The remaining columns repeat our set of three dimensional data for additional time periods (February and March). We can draw this additional dimension as being additional cubes each labelled by time period.
Our data set can be drawn as a set of cubes made up of individual data points described by a value on each of the dimensions (for example “UK Total, Sales, Actual, Jan-14”). This is illustrated in the diagram below.
In practice, we usually find there are more than four dimensions of data pertaining to any business. For example, we may also track performance by product, customer or business function and we may have results in other currencies to contend with. We can’t easily represent more than four dimensions in a diagram but the principle remains the same.
Whether we use specialist programs to create visualizations or simply the capabilities Excel, understanding the dimensions of the data we are charting is of great importance in ensuring consistent and relevant analyses.
We will look at each of the main dimensions of MI in turn in a series of upcoming posts on this blog.