Introduction
\label{introduction}
HIV vaccine studies vary in scale and complexity. Some involve a handful
of animals while others include thousands of human volunteer
participants and multiple immunoassays. A single funder may sponsor
hundreds of studies over many years, each with its own specific goals.
Centralized teams often analyze the data, which usually leads to
publication of results on the planned goals. However, investigators in
our design research indicated they have no easy way to get basic facts
or explore ideas that are not in a paper. Labs may share individual
datasets according to a funder policy but without sufficient annotation
for the benefit of others. In fact, there are reasons investigators may
prefer to keep data to themselves [1].
In 2010, the Global HIV Vaccine Enterprise published its second
strategic plan. Calling HIV/AIDS “one of humanity’s greatest
challenges,” the authors note unique properties of HIV that have
resisted development of a safe and effective vaccine since its emergence
in the early 1980s. Part of the plan calls for broad and rapid access
across dispersed study data. The authors hoped to foster a more
collaborative culture and spark diverse hypotheses through shared use of
well-annotated, harmonized data [2]. Our work set out to address
this need in the hopes of accelerating scientific progress.
This paper will discuss related work, needs that may differ from common
visual analytics (VA) research, application design, evaluation, and the
path forward.