Generating hypotheses about novel data can be tough. Without a top-down approach to explore data, even sophisticated analysts might invest hours “slicing and dicing” in dozens of R and Python charts, Excel wrangling, or using a different BI tool. Significant and tedious effort is expended just to get a feel for the data – or else risk missing relationships or biasing models.
To address those shortcomings at Einblick, and enable rapid hypothesis generation we designed a Key Driver Analysis engine (KDA). Our engine allows users to systematically evaluate thousands of sub-populations and evaluate their impact with respect to the whole population, to help our users find the “needle in the haystack” in a matter of seconds.
More specifically, Einblick’s KDA engine can be used to:
- Characterize a single population: users can study distributions within a single data population, and relate different attributes within that population.
- Find the differences between two populations: users can find patterns that are highly frequent in one population but rare in another one and vice-versa.
In sum, Einblick’s KDA represents the express lane between descriptive and predictive analytics. By finding relevant subpopulations in data and identifying candidate drivers, users will more efficiently drill down their data to find key insights.