[FoRK] Insurance coverage, Big Data/Open Data, Amanda F*cking Palmer & the U.S. Presidential Debate
bkdelong at pobox.com
Mon Oct 15 11:08:14 PDT 2012
What in gods name do any of those have in common? Well, first read this:
Boston-based musician Amanda Palmer (@amandapalmer - married to famed
UK-based author Neil Gaiman) was tweeting on a train ride back home about
an experience she had as a musician in her 20s lacking health insurance.
Her parents had watched her step-brother die - he was sick from a disease
while uninsured 0 and was fighting with her to get insurance. The just of
it is they helped her with it and she was lucky. So she conducted an
informal poll with fans on Twitter starting last night using the hashtag
#insurancepoll and it has gone wild.
Mid-morning she started trending nationally in the US, a couple hours later
- in the UK as well. She has gotten millions of tweets after asking for a
simple data set:
1) COUNTRY?! 2) profession? 3) insured? 4) if not, why not, if so, at what
cost per month (or covered by job)?
She had a statistician from a University come forward early and offer to
manually start tabulating results (@aubreyjaubrey - as the article states)
who mentioned she'd crack out a few hundred today. I quickly realized given
the second round of U.S. Presidential debates tomorrow night, this could be
a fascinating poll/study and I KNOW there have been people who have created
tools that spider the Twitter API, could pull the data down, determine
which accounts are sockpuppets and bots, segment out the non-formatted
responses for later processing, process everything else, and even decompile
URLs to determine which is spam and which isn't. Some people gave their
response in Retweets, others added additional, useful commentary.
So I put out a few calls for Big Data or Open Data companies and groups
about the opportunity. If the information manages to get normalized in the
next day before the debate, some useful statistics could come out. But the
biggest issue is truly validating accounts. How many people could spout BS
from an account they just started when perhaps a spider software could
check the accounts for # of tweets, account age, # of valid accounts that
follow it etc.
Since FoRK has the greatest amount of data scientists, enthusiasts and
people of like-mind on this list that I know, I thought I'd throw this idea
out there. I see the opportunity, I know the tools are out there, I'm
looking for them and I have the connections at the moment to possibly help
get this done in a more timely and useful manner than if just done so
manually - even if crowdsourced ad hoc. I'm not sure how much of the
Twitter Account Validation will actually be done in this case.
Thoughts? Insights? Anyone want to throw their hat in? Contact Amanda or
Aubrey directly or email me. I'd love to see this get done.
B.K. DeLong (K3GRN)
bkdelong at pobox.com
GPG Key Fingerprint: 5EEF0ABDACDD937AD08F4AF0E42DFD9081DE7CB
More information about the FoRK