Data Discovery

Extract and graphically visualize unknown patterns and relationships.

Approaching Data Discovery

Data Discovery is a DataSonar module aimed at companies that are still appraising the value of their data and focusing on business exploration.
Discovery’s main purpose is to allow a line-of-business, non-IT-expert user to explore a sub-set of data based on a business question. This is achieved in a discovery-oriented, problem-solving approach, wherein the user finds a way to translate business goals into a series of guided experiments.
To support the exploration process, the user creates a storyline of data findings based on different discovery modes.

Interactive and iterative approach

Ideally, minimum effort is required of the user by permitting an interactive and iterative approach to the exploration path. That approach includes presenting or suggesting, mostly automatically and according to the user’s standing along the path, different data, methods, algorithms, and tools.
It is significative that the discovery effort begins with exploring data that may or may not be relevant to focused analysis.
In any case, the output scenario will fall into:

  • insights.
    Defined by discoveries that change people’s perspective or understanding.
  • models.
    Defined by the creation of externalized models that describe behavior, structure or relationships in clear and quantified terms.
  • data products.
    Defined by packages of data which have utility for other analytical or business purposes.

Discovery modes

The discovery process follows Joe Lamantia’s empirical taxonomy model of search and discovery, wherein each mode describes a search-strategy goal and associated algorithms typically employed by information seekers.
The default Discovery modes are:
1) locate; 2) verify; 3) monitor; 4) compare; 5) comprehend; 6) explore; 7) analyze; 8) evaluate; 9) synthesize.

Discovery exploration patterns

Based on the taxonomy, the modes tend to cluster up and form distinct chains or patterns analogous to business needs. However, the user is free to create their own discovery modes or exploration patterns at any time.
The most frequent patterns are:
1) comparison-driven optimization;
2) exploration-driven;
3) strategic insight;
4) strategic oversight;
5) comparison-driven synthesis.

© DataSonar 2016