Share:

Data Visualizations in Business Analytics

Overview

One of the key drivers in the Pyramid 2018 design was the data visualization engine and the flexibility it affords to users in visualizing data for analytics.  This required capability needed to not only address the multitude of variations users are looking for in how they ‘draw’ their data, but also allow the application to port common visual elements from one type of visual to the next and to also support the many query styles and ‘shapes’ that a user can create in Pyramid’s Discover tool. The last issue to address related to adding more advanced analytic functionality through the visualization process – so that clever things can be used by non-technical people in an intuitive, point-and-click UI.

Click here to see an example of the many permutations that can be created in Pyramid.

Drop Zones to the Rescue

The solution in Pyramid 2018 was ultimately driven through the drop zone engine which was designed to take the old “columns” and “rows” concept - used over the last 20-25 years in products like Excel, ProClarity and even Pyramid’s own BI Office – to the next level. And although there are similar concepts in other products like Power BI, Spotfire and Tableau, the Pyramid 2018 version is designed to paradoxically both simplify the functionality and extend its sophistication at the same time. Much of this can already be seen in the current Pyramid release; with more of that to come.

Core, Extended and Custom Visualizations

Currently there are 36 core data visuals or charts available in Pyramid (with more on the way). Using the drop zones, these can be extended to well over 150 differentiated visualizations that do not include permutations for formatting or ‘soft’ adjustments in the design of the presentation (soft includes headers, axes, legends etc.). And to further fuel this list, Pyramid will soon be adding a custom visualization API and programming model - based on the D3 JavaScript graphics library – potentially extending the capability to an endless list of visualization options.

Walk-through Example

Watch the video

The following walk through will highlight the drop zone mechanics in Pyramid 2018 and demonstrate the flexibility and power of Pyramid’s data visualization engine. Using a standard column chart setup, we are able to create 12+ extended variations with some basic changes in the drop zone configurations.

Start by opening a new Discover session using the In-memory Sample Demo database provided in the standard Pyramid install.

Sample Demo Database 

Add the manufacturer hierarchy to a column chart and then click Sales in measures.

Data Measures

You’ll get the chart shown below – a standard column chart - with manufacturers as the categories in the chart.

example column chart 

Dragging the manufacturer blue chip to the color drop zone, will instead show manufacturers as the series in the chart, each with its own color.

Color drop zone

Now drag the Gender hierarchy (column) from the ‘customerPorfile’ dimension (table) to Color. This will provide a dissection of Sales by Manufacturer by Gender.

Vertical Trellis

Moving the Gender blue chip to the Vertical Trellis will create a multi-chart of Manufacturer Sales, broken out by Gender.

horizontal trellis

Moving it to the horizontal Trellis has a similar effect. Be sure to choose ‘Scale to Fit’ in the component formatting ribbon to see all charts at once.

Gender Horizontal Trellis

By adding the Gender hierarchy to Color as well as Trellis, we are able to color each chart series in the multi-chart by the same elements driving the multi-chart.

Multi Chart

Let’s Move back to a Vertical Trellis, and add the measure ‘Net Profit’ to The Size Drop Zone. This creates an interesting effect where each bar’s thickness is driven by net profit, while its height (the usual column chart effect) is driven by Sales.

Stacked Column Charts

Right click on the Column Chart widget in the toolbox, and choose Stacked Column Charts. And Remove the Gender chip in Vertical Trellis.

Size Drop Zone

This will change the columns into a stacked column chart, with bars broken out by Gender.

Column Chart widget

Moving ‘the orange ‘Net Profit’ chip to size, will again produce an interesting chart, where the width of each data point in the chart is driven by net Profit and its height is driven by Sales.

bar chart

Next, remove Gender, and drag Net profit to the color chip.

Color drop zone header

The columns in the chart now are colored using a simple linear model of values from high to low based on Net Profit figures for each manufacturer (The height is still being driven by Sales).

Next, drag the Net Profit orange chip back on to the Color drop zone header, and then choose the ‘Average’ Sub-drop zone in the fly out.

data visualization

The columns are now colored using above or below average Net Profit values.

Values drop zone

Next, drag net profit to the Values drop zone.

Bi viz

 Since there are 2 values, the engine automatically builds a special “Values” orange chip and puts it into Vertical Trellis – to show both measures.

Business Intel Visualizations
Redrag Net Profit onto the Values drop zone header, and choose Secondary Axis > Spline Chart.


Secondary Axis

And now you have a Combo chart with both Sales and Net Profit combined.

Spline Chart