Seth grimes bio:
Seth specializes in strategic
IT analysis, architecture &
planning with a focus on
business intelligence (BI) and
text analytics systems.
Seth consults via Alta Plana
Corporation, organizes
the Sentiment Analysis
Symposium, and writes for
InformationWeek and other
UBM TechWeb publications.
For consulting logo-porn, visit
http://www.altaplana.com/
clients.html .
For additional background,
visit http://sethgrimes.com
or follow
http://twitter.com/SethGrimes
Is social media with its
obvious
noise
a
right
platform to collect opinions?
Does this new fad of social
media
research
really
effective?
Seth Grimes
employing
proven
methods
where possible and assessing
results reliability and significance
in all cases, conducted with
respect for transparency and
repeatability. The client should
insist on these factors. If the
client does, and the researcher
adheres to them, the client can
focus on what she does best:
ensuring that research answers
business needs.
Transformations
performed
including
the
approach
to
semantic
labeling.
Sample
design. Tested, computed results
significance. These are factors
in data quality assessment
and preservation. You won’t
accurately know what you
have – you won’t be able to
completely judge the validity
and reliability of derived insights
– if you don’t respect these
factors.
What are some of the data
quality parameters that How can the researchers
clients cannot afford to maintain balance between
quality and timelines?
ignore?
Seth Grimes: Data provenance
and context. Extraction method.
Seth Grimes: Planning.
Seth Grimes: Actually, there is
very little noise in social media. I
think the concept you want isn’t
noise, its salience. Very little of
what’s out there is relevant or
needed for a given study. Very
little is salient. But all the rest –
that’s relevant to something,
even if not immediately useful to
you. Salient social data carries
immense value as an opinion
source, and you can tap the data
on-demand, without time and
expense involved in conducting
a survey or convening a focus
group. So yes, social is the right
platform for opinion mining, so
long as you can find the right
data and find analysis methods
that produce the insights you
need. And yes, research that
draws on social media (and also
social-media research, which is a
different species), can be really
effective. so long as you can find
the right data and find analysis