Twitter as a Predictor
October 27, 2011
Rhyme and alliteration accompany Twitter. Example: Twitter trending topics are often a big hit or miss when it comes to reflecting evolving events, such as the Occupy Wall Street movement.
However, that isn’t stopping an emerging industry aimed at using tweets of millions of people to help predict the future. How is this possible? By following certain terms surrounding everything from disease, elections, and finance, you can gain insight into what may happen. For example, during the Egyptian revolution earlier this year, there was a high correlation between tweets and actual events. There is even a hedge fund called Derwent Capital Markets that makes stock and fund trades based on Twitter analysis, and it is actually fairing well.
“Can Watching Twitter Trends Help Predict the Future?” on GigaOM tells us more:
The theory behind all of this Twitter-mining is that the network has become such a large-scale, real-time information delivery system (handling more than a quarter of a billion messages every day, according to CEO Dick Costolo at the recent Web 2.0 conference) that it should be possible to analyze those tweets and find patterns that produce some kind of collective intelligence about a topic. It’s the same idea that drives companies to do “data mining” on their customers’ behavior…
Will this become a “must have” tool for researchers, medical staff, and politicians. Even the U.S. government’s Intelligence Advanced Research Projects Activity Unit is looking into using data from social media as part of its intelligence gathering. But is passivity better than active research? We think one needs both. Judgment helps too.
Andrea Hayden, October 27, 2011
Sponsored by Pandia.com
Comments
2 Responses to “Twitter as a Predictor”
-
- Subscribe to Beyond Search
Feature archive
News archive
-
Following ‘Trending Topics’ occasionally indicates trends “on the edge of the radar screen so to speak!”
Trend-spotting is more of an art than a science; that is even the most advanced algorithms applied by social-network sites, and search-engines alike still discern potential trends through scientific approaches.
One method for trend-spotting (as more of an art, and less of a science) is called inference scanning; that is the scanning of blog and news content containing words potentially indicative of emerging trends. Yet, even inference scanning cannot always discern the contexts of just what constitutes relevant trends. One advantage of ‘inference scanning’ are the narrowed-down listings of search-engine results.
A few (of many) examples of emerging-trend keywords are: ‘TREND’ ‘BECOMING’ ‘INCREASINGLY’ ‘LATELY’ and ‘MAY.’
Quite a few search-results can indicate those important trends which otherwise would “get lost in the shuffle so to speak.” A few of the search-results may even contain content which is “the closest thing to a crystal-ball.”
Four search examples (LINKS) contain combinations of the words ‘TREND’ ‘BECOMING’ ‘INCREASINGLY’ ‘LATELY’ and ‘MAY.’ For the most up-to-date content, search-results list blog or news content over the previous 24 hours:
* Contains words ‘MAY’ AND ‘TREND’ – Contains any one or more of these words: ‘BECOMING’ ‘INCREASINGLY’ ‘LATELY’
http://goo.gl/AXWPP
* Contains words ‘MAY’ AND ‘BECOMING’ – Contains any one or more of these words: ‘TREND’ ‘INCREASINGLY’ ‘LATELY’
http://goo.gl/MqnLU
* Contains words ‘MAY’ AND ‘INCREASINGLY’– Contains any one or more of these words: ‘TREND’ ‘BECOMING’ ‘LATELY’
http://goo.gl/G4Zuf
* Contains words ‘MAY’ AND ‘LATELY’—Contains any one or more of these words: ‘TREND’ ‘INCREASINGLY’ ‘BECOMING’
http://goo.gl/EVAEt
FURTHER DETAILS:
Inference scanning also involves rejecting content containing such words as: ‘HE’ ‘SHE’ ‘YOU’ ‘I’ ‘ME’ – as mentions of “he said” “she said” are often indicative of individual statements, which are usually not representative of collective thinking and actions associated with emerging trends.
For the purpose of narrowing-down search engine results pages as much as possible, the term ‘MAY’ can apply for all searches (the search-engine also provides variations of ‘MAY’ such as content containing the terms ‘CAN’ and ‘COULD’). Content containing the terms ‘MAY’ ‘CAN’ ‘COULD’ are favorable terms towards inferring relevant and emerging trends in many diverse topics.
Michael Ross,
Thanks for the post. We don’t do news or trends. We opine. See http://www.arnoldit.com/wordpress/about
Stephen E Arnold, October 29, 2011