Tutorials > Chart Tutorials > Small Multiples

Small Multiples

One of the best innovations in Tufte’s “The Visual Display of Quantitative Information” is the idea of using an array of miniature charts to display multi-dimensional data. Often a series of values are recorded at (say) several sites or on several days, then are superimposed on a single chart as a collection of dotted and dashed lines in various colors. This often makes it very hard to see the patterns and relationships in the data.

The ‘Multiple’ plots each series on its own chart, and arranges them automatically in a trellis to allow easy comparison between them.

Comparing two datasets

This example is the simplest form of multiple chart. Two sets of data are compared by being plotted side-by-side as barcharts. Note that the multiple chart has very few styles of its own – here the example uses the YAxisStyle to have both charts constrained to show the zero point.

sp.SetPageLabels(new string[]{"1995","2005"});
sp.SetXLabels(new string[]{"Milk","Eggs","Bread"});

sp.YAxisStyle = YAxisStyles.ForceZero;
sp.SetLabelFont("Arial",13);
sp.SetCaptionFont("Arial",15,FontStyle.Bold,Color.Firebrick);

sp.DrawMultiple(ChartType.BarChart,data);

PageLabels are a useful way of labelling the cells in the trellis.

A Collection of Histograms

The barley-yield data could be shown on a single histogram, but often this conceals some systematic variation in the data. In this example, the farm is used to create a trellis of charts, each showing the spread of the data at the individual farms.

Although the experiment was designed to investigate various barley varieties, this plot makes it very clear that the farm is going to be a big factor in the outcome. A quick glance shows that yields at Wasaca may be quite variable, but that it easily beats both Grand Rapids and Duluth (which is consistently very poor).

sp.SplitBy(_barleyfarm);  // One cell per farm

sp.HistogramStyle = HistogramStyles.SurfaceShading;
sp.ClassInterval = 10;
sp.SetXRange(0,70);
sp.SetYRange(14);

sp.SetCaptionFont("Arial",14,FontStyle.Bold,Color.Firebrick);
sp.DrawMultiple(ChartType.Histogram,_barleydata);

The PageLabels have been automatically set to reflect the unique values in the dimension used to split the data.

Constructing a dashboard

This example automates a typical use of the dial chart, which are often grouped in arrays to show related values.

sp.SetXRange(100);
sp.SetXZones(new Zone(80,100,Color.Red,FillStyle.Solid));
sp.SetPenWidths(1.5);
sp.DialChartStyle = DialChartStyles.ArrowLines;

sp.SetPageLabels(new string[]{"Engine No.1","Engine No.2","Engine No.3","Engine No.4"});
sp.XCaption = "Coolant °C";

sp.SetLabelFont("Arial",13);
sp.SetCaptionFont("Arial",15,FontStyle.Bold,Color.Firebrick);

sp.DrawMultiple(ChartType.DialChart,new int[]{23,34,78,12});

Note that the range and warning zone will be common to all the charts in the trellis.

Comparing several attributes

The final example shows a collection of polar charts, which readily allow various criteria to be compared across several cameras (the example came from a Swiss camera magazine). Notice how the distinct set of camera types is automatically used as page-labels to caption each chart in the trellis.

sp.SetMargins(36,12,20,0);

cameradata = new double[] {4.5,3.1,2,4.5,5,2.5,5,5,3,5,3.5,2.5,4.5,3,3,4.5,2,4,4.5,
         4.5,2.5,4.5,4.5,4.5};
cameratype = AE.Compress(6,new string[]{"Sony\nCyber-Shot\nDSC-V1",
         "Canon\nPowerShot\nG5","Nikon\nCoolPix\n5400","Olympus\nCamedia\nC-5060"});
cameracap = new string[]{"Résolution","Rendu","Couleur","Sensibilité","Flash",
         "Autofocus","Rapidité"};

sp.SplitBy(cameratype);  // Sets trellis

sp.SetFillStyles(FillStyle.Opacity42);
sp.SetColors(Color.Red);
sp.SetMarkers(Marker.SmallBullet);
sp.SetLineStyles(LineStyle.Solid);
sp.SetPenWidths(1);

sp.PolarChartStyle = PolarChartStyles.SurfaceShading|PolarChartStyles.GridLines|
         PolarChartStyles.Markers;
sp.YAxisStyle = YAxisStyles.ForceZero;
sp.SetXLabels(cameracap);
sp.XLabelFormat = "XXXXXXXX;";
sp.SetLabelFont("Arial",6,Color.Black);
sp.SetCaptionFont("Arial",14,FontStyle.Bold,Color.Black);

sp.DrawMultiple(ChartType.PolarChart,cameradata);

This is likely to make visual comparison much easier than attempting to overlay the radar plots from several cameras on the same chart.

Send comments on this topic
© Dyalog Ltd 2013