Matthew Beckman (Penn State University), Christopher Desjardins (Saint Michael's College), & Lisa Lendway (Macalester College)
Abstract
Exploratory Data Analysis (EDA) perhaps suffers a wider gap between real & perceived value than any contribution the data analyst offers. The popularity of Data Science has ushered in a rising tide of open-data initiatives and elegant software solutions readily accessible to non-specialists, thereby democratizing data analysis for all manner of subjects. Initially, some novices may feel overwhelmed and others may risk spurious results utilizing opaque methods with fancy names, yet all can be empowered by careful EDA based on lucid methods powered by curiosity, creativity, and critical thinking to build fruitful intuition and responsible insights. However, EDA too often seems a cursory obligation to be minimized.
The session will include several focus group style conversations in small groups reflecting on issues including the evolution of contemporary EDA, top priorities & best practices for a careful EDA to be reinforced with our students, and other topics. As a result, the session aims to highlight areas of consensus and debate as we (1) recast/affirm goals and considerations for EDA in a world laden with found data, big data, etc. (2) rethink where/how EDA should be addressed in data analysis courses at all levels.