Mining Twitter
Lots of reports – mining and tagging twitter data
Backtweets
obsessed with tool – has not been demoed – cloud.li – the ideal, color the items in the cloud, red green or grey, then hoover brings up tweets, and you can see, slider over time, dream dashboard
capture and interpretation – tons of sentiment tools;
search.twitter.com – sumize for detecting wether the text is positive or negatively, sentiment analysis to twitter analysis, sentiment score in the apis
– tools: https://twendz.waggeneredstrom.com/
Stock mood – one of the techcrunch 50s
many have tried to get sentiment right,
as a brand manager, look
Tim Orilley – industry ranked hot lists, of brands
Celebrities –
people browsr – negative tweets, 1 cent to pass through special interface, stephanie, adventure girl, travel hot list.
going public very soon, palm pre – twitter api hicuup, what the api says,
Algorithyms
rooms full of people
KT Payne – rooms full of people to analyze
reports on mac book air – bubble routine
ignore nuetrals – high level, features, most discussed. battery big negative, life was not good
Exports: twittergroups
track a link, turn a stack into a rss
backtweets.com
myinview.com – Twitturly inView is a powerful solution for monitoring, receiving instant notifications, and responding to anything that matters to you on Twitter. We are currently in Private Beta, so please request access below if you want to see what we have to offer.
tweetbeep.com – email alerts of terms that match on twitter
Trendistic – type in any term
https://trendistic.com/ – public timeline, great for trends, not helping
Make software tools. Check out twist.flaptor.com for twitter trend charts.
500 posts a second, 4 billion twits a year, could be 40-50 billion in a year, 3 trillion sms messages per year
https://search.peoplebrowsr.com/ monday they will have 30 days data
3 different
streaming, search and rest, apis
1. search.twitter.com
good-
bad – doesn’t go back very far about 3 weeks, you can’t go back
suck ing data through the RSS,
giant spreasheet, lots of rows, you get everything via the feed, very few columns, three, date, person who wrote and the tweet
want more but don’t want to pay for it
search.peoplebrowser.com – export to csv
next week rolls out 30 and 90 days search
sentiment
analysis core data
2. tap in through the API
Joel – twitterly
product called urltrends
Jody
WingDude on twitter
BJ Fog – facebook applications
Mining Twitter: Extracting Value From Conversations and Connections
When people talk about promising uses for the rich data stream provided by the Twitter APIs, discussion frequently turns to two related applications: Brand monitoring and investment management.
In both cases professionals with discretionary budgets desire timely (and polar) commentary that surrounds products and companies. In addition, these two arenas often require managers to take immediate and appropriate action with limited information. The right decisions can build fortunes, the wrong ones can doom careers (and companies.)
Twitter data is uniquely suited to these applications. The immediacy of the information flow, along with the ability to detect geography and relative influence is unparalleled. In this session, we’ll see how leaders in both spaces are taking advantage of what Twitter has to offer.
Dealing with sentiment * Assigning weight and importance to extracted data * Structuring: Filtering and tagging methods.
