A small group of us (mainly Harvard students, but others as well) had
similar questions a few months back and we've started digging into the
research pretty heavily (also doing contracting/consulting). Our group
is called the Web Ecology Project and we've released a handful of
academic (but accessible and market relevant) papers about events and
influence on Twitter. We also released some source code for
identifying the language of Tweets using the Google API and python.

We have a research relations arm and I've forwarded your contact
information to our research coordinator Dharmishta who will probably
be in touch soon.

It might be worth looking over our papers. When you do, feel free to
email me (or any of us) and ask questions. We've done some semantic
analysis of tweets for our paper on the death of michael jackson and
the response on Twitter using the ANEW method/dataset and NLTK.

Our work can be found at http://webecologyproject.org and all of our
stuff is released under a Creative Commons license so feel free to
quote us and use bits where you need to.

One of the hardest initial things that we encountered with Twitter was
gathering and storing of useful and meaningful data but now we've
gotten mostly past those issues and we are now starting to mine and
analyze other social networks as well.

Thanks,

David Fisher
Web Ecology Project

On Sep 27, 8:11 pm, Stefna <[email protected]> wrote:
> Thank you all for the feedback.
>
> My main motivation for posting this thread was to gather some loose
> ideas. One man’s trash is another man’s treasure, posting a link costs
> nothing but saved a lot of mine time.
>
> "strict formed data" - 140 chars, #tag, @username, RT etc. - that's
> why there are so many sites presenting graphs, charts, trends,
> tendencies etc.
>
> "first pick was vaguely to analyse the semantic meaning of tweets" -
> like take a tag, take all the words with tweets with this tag and on
> that base find tweets that might me related to the topic. Or finding
> and measuring tags that are in tweets with aforementioned one.
>
> I should keep in mind that my task is to present a satisfactory
> dissertation, not cure for cancer. And probably by the time I finish
> my work someone else will ship similar software independently.
> Nevertheless by creating it I will learn a lot of stuff about Twitter
> itself not to mention improving programming skills.
>
> On 27 Wrz, 20:38, Mitchel Berberich <[email protected]> wrote:
>
> > Careful! Stefna was talking about semantic meaning. Not
> > "syntactic" ...
> > But I think you're right - Stefna, please us tell a bit more about the
> > context.
> > And - what do you think of when you say "strict formed data"?
> > What exactly do you want to achieve?
> > Maybe your promoter should tell you in more detail, what he expects
> > from you?
>
> > On Sep 27, 6:44 am, Nalin Savara <[email protected]> wrote:
>
> > > Good luck buddy.. Btw I'm curious, What exactly are you referring to
> > > or focussing on when you say 'syntactic meaning of tweets' ?
>
> > > I mean I'd appreciate a clarification on 'syntactic meaning to whom ?'
> > > and 'syntactic meaning of tweets in what context ?'
>
> > > Just wondering..
>
> > > Best Regards,
> > > Nalin
>
> > > On 9/25/09, Stefna <[email protected]> wrote:
>
> > > > I've submitted a ticket with following content:
> > > > *** *** ***
> > > > I am a 23 years old student of informatics at AGH Universtity of
> > > > Science and Technology in Cracow (Poland). Due to a rapid development,
> > > > strict formed data and accessible API I would like to designate my
> > > > master thesis to the Twitter related topic. My promoter is the PhD at
> > > > the Department of Computer Linguistics and our first pick was vaguely
> > > > to analyse the semantic meaning of tweets.
>
> > > > Do you have suggestions about the dissertation topic?
> > > > Do you have any pending requests or prospect features you want to
> > > > develop?
>
> > > > I will browse known issues, I will think thoroughly about the topic
> > > > but still - your suggestion might be very helpful. Even the shortest
> > > > one (like "good luck") will encourage me to more intensive research.
> > > > *** *** ***
> > > > Does anyone have any suggestions? My ticket has a six-digit number so
> > > > I'm afraid I won't get any answer :)
>
> > > > I'll probably ask for help during my work so I subscribe to this group
> > > > anyway.
>
> > > > Thanks in advance!
>
> > > --
> > > Sent from Gmail for mobile | mobile.google.com- Hide quoted text -
>
> > > - Show quoted text -

Reply via email to