Browse Prior Art Database

Methods to visualize extra related dimensional data in heat maps Disclosure Number: IPCOM000244130D
Publication Date: 2015-Nov-12
Document File: 4 page(s) / 68K

Publishing Venue

The Prior Art Database


Described are ways of using visualizations that have the quick insight of a heat map but allow additional dimensions of data to be analyzed at a quick glance. Heat maps allow you to quickly analyze data by showing a series of colored cells. If your data has multiple dimensions that fit within a cell, you can collapse to a single color through some algorithm; however, this is a lossy transformation.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 52% of the total text.

Page 01 of 4

Methods to visualize extra related dimensional data in heat maps

A typical heat map is implemented as a two-dimensional grid of color intensities. The example heat map below uses the intensities to represent the impact of a side effect on a patient as a result of using a drug (impact here is a function of the probability of the adverse effect occurring and the severity of occurrence). Heat maps can be used to compare some value across two dimensions, the example looks across columns to see which treatments are most toxic, and look across rows to see which side effects are most common.

    However, a typical heat map forces us to come up with a single value for each box, and in this use case collapsing to a single value causes a loss of useful information. There is a need to be able to show more dimensions of data, yet still have the "quick comparison" benefits of the heat map.

    The core idea of this invention is to convey multi-dimension information from a cognitive calculation into a visualization where the dimensional information are viewable within a specific category, covering three dimensions of data simultaneously across multiple potential answers. Two related dimensions are combined into a cognitive score based on the data points, then the base values are normalized against the set of values, then utilized as a percentage against a bar within a cell, depicting the relative values for the dimensions.

    This invention extends a traditional heat map to use additional related dimensions by manipulating the properties of the box itself. Rather than selecting only one variable for each box, or smashing the variables together in a single score, manipulate additional properties of the box. Instead of creating an "intensity score" from incidence and severity to use as color, show each via the dimensions of each box (incidence on X, severity on Y). Alternatives and problems with those alternatives:

Pick only one variable, or smash multiple variables into one: loses information

Use text overlays on each cell to provide extra information: difficult to read, very noisy visually

Use hovertext overlays on each cell to provide extra information: difficult to interact with, slow to convey additional information

Reliance on color: can be non-accessible

Key Novelty:

A calculated score across a related set of dimensions, then a normalization of the data points such that they represent a relative distance, then plot it across a heat map in the form of a bar chart, within a cell.


For the sake of the following examples imagine the following table.
Cisplatin/Docetaxel Paclitaxel