I’m a graduate student in the EECS Department, working with Professor Marti
Hearst in the School of Information, at the University of California,
Berkeley.

We are conducting a research study to understand how experts conduct
exploratory data analysis (EDA). We are recruiting professionals who
analyze data as a major part of their daily job to interview about their
approach to EDA, including topics such as workflow, tools used, type of
data analyzed, and techniques employed. Please see below for additional
background.

If you regularly analyze data as part of your job and would like to
participate in this research, please email alspa...@eecs.berkeley.edu to
schedule an interview or ask any questions you may have.

The interview will be conducted at a time and location of your choice. It
will last one to two hours and you will be compensated $25 per hour in
Amazon gift certificates for your participation.

We hope that the information gained from the study will help data analysts
and researchers learn about current the best practices, methodology, and
day-to-day workflow of expert analysts. We really hope you will choose to
participate!


Thank you for your time,

Sara Alspaugh


*Additional Background*

EDA is an approach to analyzing data, usually undertaken at the beginning
of an analysis, to familiarize oneself with a dataset. Typical goals are to
suggest hypothesis, assess assumptions, and support selection of further
tools, techniques, and datasets. Despite being a necessary part of any
analysis, it remains a nebulous art, that is defined by an attitude and a
collection of techniques, rather than a systematic methodology.

Because of this, I argue, exploratory analysis is not well-supported by
current analysis tools. Better understanding of how exploratory analysis is
done in practice would allow us to develop tools to support important
goals, such as (1) ensuring complete analysis coverage of a data set, (2)
allowing different researchers working independently to come to the same
result, (3) dividing the burden of exploring a data set evenly among
collaborators, (4) managing complex and long-term analysis workflows in an
orderly fashion, and (5) automating that which can be automated.

This will be a qualitative study that seeks to gather information about
current the best practices, methodology, and day-to-day workflow of expert
analysts, particularly with respect to EDA. This information will help us
develop systematic yet flexible methodological frameworks and software
tools to better support people conducting EDA.
--
The list archive *is* the website
https://www.mail-archive.com/socialcomputing@lists.berkeley.edu/

Subscribe on the web!
https://calmail.berkeley.edu/manage/list/listinfo/socialcomputing@lists.berkeley.edu

Subscribe with email!
mailto:socialcomputing-requ...@lists.berkeley.edu?subject=subscribe

Reply via email to