DDMA Mining WebEx 4/13/12

siberiaskeinData Management

Nov 20, 2013 (3 years and 11 months ago)

219 views

DDMA Mining WebEx 4/13/12


Rahul Ramachandran began discussion to determine how we drill down to
determine the best areas to concentrate in order to improve data mining
capabilities. He said we could put out a survey to obtain inputs.


Joseph Baker put
together discussion questions; recommends that we get
geoscientists to talk to data mining experts to explain what they need; with a goal of
developing mining services that will enhance productivity, with user friendly
interfaces, etc.


Chris
Lynnes
chat c
omment, “
This reminds me of the 2nd NASA Data Mining
Workshop in Pasadena, 2006, very similar questions...I wonder how far we have
come...”


Hook Hua suggested that we make data mining approach driven by science needs.


Anne Wilson stated that she thinks s
cientists have survey fatigue; that several other
groups are already using surveys. She mentioned that we should consider Domain
Specific Software Architecture.


Chaitan chat comment, “
NSF is calling a meeting next week of workshop PIs. we
can discuss the

survey issue there.”


Erin Robinson agreed with survey fatigue, but suggested we could coordinate with
other groups (high level coordination)

and have a combined survey
; perhaps
with
brokering/workflow.


Anne Wilson thought coordination with other group
s would be too difficult in this
timeframe.


Chris Lynnes said if we use surveys we should limit the number of questions.


Ruth Duerr mentioned that
the Brokering
group is

using the concept of liaisons
.

Jay
Pearlman is
the Brokering
liaison to data mining
. A couple of surveys are already
out; but don’t know if they have much relevant to data mining. Their survey will be
posted to the ning site.


Rahul asked whether we should go forward with
the
survey or not.


Chris stated that we should proceed cautious
ly.


Joseph noted that there are different types of e
xperts; for example, he

is a user.

The
questions should be geared toward the type of user.


Ruth chat comments, “
Survey for data providers:
http://www.surveygizmo.com/s3/880349/EarthCube
-
Broker
-
Data
-
Provider
-
Survey
” and



Survey for data users:
http://www.surveygizmo.com/s3/883858/EarthCube
-
Broker
-
Data
-
User
-
Survey



Anne suggested we table this discussion until after her presentation on Domain
Specific Software Architecture.


Chaitan Baru said we could ask workflow and on
tologies groups about coordination.


Hook stated that surveys can help to determine state
-
of
-
the
-
art.


Chris pointed out that we could also review literature.


Stefan agreed that a literature review would be good; th
is group could take a stab at
it from
the literature
.


Lee Allison chat comment,

“The Governance Roadmap group is summarizing
governance content from all the white papers and EoIs and found them to be
very helpful in shaping our process and identifying issues.”


Chaitan chat comment,
“NSF is c
alling a meeting next week of workshop PIs. we
can discuss the survey issue there.”


Joseph

mentioned that use cases are important.


Steve Richard reminded us that the deliverable for this effort is a roadmap.


Anne Wilson presented “Towards a DDMA
Reference Architecture,” including a
poster showing flow.


Joseph asked about the heritage of DSSA.

Ann stated that the
DSSA
process was
developed in the 90’s by the Software Engineering Institute.


Joseph asked what projects it had b
een applied to. Ann
was unsure, but noted that
there were a number of similar processes.


Joseph asked if she thought it would be possible for folks to come up to speed on the
process and accomplish our goals within the stated timeframe.


Chaitan stated that it would be good
to interact with ontologies/interoperability. He
also noted that scientists are interested in space and time, but other topics bubbling
up to the top includes data quality and data use. One thing that would be nice is if
the t
ools could make suggestions
from what
data
is available



Anne mentioned the importance of metadata.


Chaitan mentioned that there are too many options to support everything and that
we will need interaction with other groups.


Ann
e

asked how we could dovetail with other efforts. We

need to identify
stakeholders, data users, data, then identify representatives, work on domain model
vocabulary, etc. to determine what functionality we need.


Ruth asked what is the domain


Chaitan mentioned that there should be posting on the ning site
about coordinating
software engineering across EarthCube
.


Tanu Malik said that Jim Boling’s EOI addressed how the project is conceptualized,
also someone from University of Miami had a similar topic is their EOI.


Anne stated she didn’t think we could do
this in the timeframe required and that it
would probably take 6 months, but we could just start and see how far we get.


--
end of Mining WebEx