SBIR Phase I: The Data Context Map

Period of Performance: 07/15/2017 - 06/30/2018

$225K

Phase 1 SBIR

Recipient Firm

Akai Kaeru, LLC
302 E 88th Str. #4F
New York, NY 10128
Firm POC, Principal Investigator

Abstract

The broader impact/commercial potential of this Small Business Innovation Research (SBIR) Phase I project is that it makes high-dimensional data visualization more accessible to mainstream users. An interactive multivariate visualization framework will assist users in the understanding of complex phenomena and help with decision making in the presence of multiple factors. It uses a map-like layout of factors and data that naturally appeals to a mainstream user's innate visual literacy, making it easy to learn and use. Through a variety of interaction facilities, users can define areas in this layout that indicate parameter values of their interest. This then enables them to make informed decisions within a single visual interface and balance among the diverse impacts the various factors have on the decisions. Given the massive growth in the availability of multivariate data, the unique capabilities that the proposed framework provides are expected to have a strong impact on society, in many application domains and at many different levels - personal, business, finance, scientific, medicine, and so on. In short, this project will have a significant societal impact by allowing users to make better decisions in less time. This Small Business Innovation Research (SBIR) Phase I project makes several intellectual and scientific contributions. A core contribution is the ability to embed high dimensional points and factors into a 2D plane or map, such that the distance between points has contextual meaning. This allows a single visualization to convey what currently takes a dashboard of several bivariate plots to show. The inherent mechanisms that create the combined map-like layout of data and factors are novel and open new opportunities to understand multivariate data. The same is true for the interactions defined on the layout, in particular those that allow users to derive the areas of influence of the various factors. Another intellectual contribution is the interface that allows users to focus their attention on specific attributes and deals with the mental and cognitive overload that may arise in users in the presence of too many factors and attributes. The proposed system constructs a meaningful hierarchical representation of the attributes that can be navigated within a novel space efficient visual interface. Using our design tools, anyone with an internet connection and a modern web browser will be able to create these layouts from multivariate data.