Visualization for drawing smart insights from Multi-Feature BigData Dataset
Publication Date: 2014-May-23
The IP.com Prior Art Database
Progressively with the enhancements in Big Data technologies, the feature sets in a given dataset from which insights need to be derived are becoming extensively huge. With such enhancements, most of the visualizations made for conventional/ non - Big Data datasets are losing their potency in delivering optimal value insights easily from the enhanced features in these datasets. The article deals with newer visualization forms and techniques that can be used to extract smart insights from a multi-feature Big Data dataset. The current scope of the article is limited to enhancing the number of distinct continuous and/ or nominal features associated with the visualization of a conventional bubble/ scatter plot.
Page 01 of 5
Visualization for drawing smart insights from Multi -Feature BigData Dataset
Background: Problem Definition, Existing Solution and its limitation.
An increasing need is felt to drive more and more insights from the visualization of data. In this regard multiple newer types of visualization are used, or the existing ones are used.
One of the very popular chart types in business and other uses is the scatter plot, where the
position of the bubbles on the 2D/3D plane depicts the relative performance of the entity represented by the bubble as compared on the attributes depicted by the different axes of the
Some more insightful flavors of the bubble chart involves the varying bubble size with a fact (say the greater the sales depicted by the bubble, the greater the diameter of the bubble, and/or the varying color (discrete or continuous gradient) of the bubbles.
But as the bubble chart is a very important and informative graph type used in business analysis, and mainly used in the scenario of multiple discrete entities represented by the bubble, a need if often felt to expand the breadth of depth/ dimensions that can be depicted by the bubbles on the bubble chart.
Page 02 of 5
Page 03 of 5
Figure 1 - Limitations of the conventional visualization.
Disclosed is a technique wherein, keeping the minor axis constant - c (say mean or min of the
proportionally weighted fact used for major axis), the additional fact (in addition to the fact used for varying color, or size, or both, or none of the bubbles) is used to vary the size of the major axis of the resulting ellipse as per the equation:
x*x/a*a + y*y/c*c = 1
x and y are the ordinates of the centroid of the ellipse on the x and the y axis
c is the standardized constant used for statistically weighing the minor axis (can be 1 or any other appropriate number)
a is a positive fa...