Quantitative Finance Stack Exchange is a question and answer site for finance professionals and academics. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Can social media sites, like Twitter, be used to analyze financial markets for algorithmic trading? How much research has been done on this topic?

share|improve this question
You can also use Google searches for that. Visit <a href="www.freakonomics.com/2011/10/07/can-google-searches-predict-stock-price-pe‌​rformance/"> this page </a> for more details. – wburzyns Nov 3 '11 at 11:27
Certainly there is a great deal of raw information available on Twitter, et al, but turning that information into something actionable is quite another manner. I'd start with a PHD related to NLP. – SCVirus Nov 12 '11 at 23:08
@SCVirus Thanks, actually I'm doing a PhD in NLP. Unfortunately, a moderator decided to delete my account.... – siamii Nov 18 '11 at 4:17
Maybe you can extend your question with asking: how effective is it expected to be? – JohnAndrews Jul 31 '13 at 16:00

Personally, I am very skeptical of the claims in "Twitter mood predicts the stock market". There are several other papers with similar claims, but not so much good quality research is available. Arguably, the sweet bits of these approaches are not public. A sounder approach is to dig at the relationship between social media activity and relate it to the stock market. One idea is take social media as a proxy for investors' attention and exploit this channel to say something sensible about the stock market.

The typical viewpoint is that social media mining provides info on the interests of individual investors, in contrast to big investment firms. Within this framework the hypothesis of Barber and Odean (2008, RFS, "Do retail trades move markets?") makes a lot of sense: retail investors buy popular and news-making stocks, pushing prices up. Interestingly, their investing activity is asymmetric: buy news-makers, but only sell what they own - rarely on news.

Some fresh good work on the topic, based on google searches, is the paper "In search of attention" by Da, Engelberg, and Gao (2011, JF) (EDIT: this is the same work pointed out by wburzyns in his comment above).

share|improve this answer

I assume that by "how much research" you mean "could you provide me with some links"....

So, as @Shane mentioned in its comment, a hedge fund recently started and is focusing on twitter analysis (here is another link).

From what I understand they are basically implementing a trend-following strategy based on the "mood" of twitter users (I, of course, don't know the details, but it's surely more complicated than that).

As for the links, we already discussed the subject of speech recognition in trading in this post.

share|improve this answer
This is correct, they also claim alpha in their strategy. – strimp099 Nov 3 '11 at 16:08

The original paper on using Twitter to predict the Dow can be found here. The hedge fund the authors are working with on trading is called Derwent Capital. An interview with the gentleman in charge can be found here. A quick Googling for "twitter analysis paper" turns up a number of results, although I am not going to post links as I have not reviewed them.

The search term factors for a model would be very interesting but very few people have access to the data in sufficient granularity for it to work. You could try licensing data from network providers such as Comcast or Verizon.

Most of the literature on traditional text mining will be applicable to Twitter although some work will be needed to normalize for document length and the unique vocabulary on Twitter.

share|improve this answer
This is brilliant advertising for Derwent (hedge funds typically cannot advertise in the US), but the original paper by Bollen et al (now famous for being famous) is laughably inadequate. (just plot the empirical distribution of the 49 p-values they quote, for example.) – shabbychef Nov 8 '11 at 22:04
I don't know about brilliant advertising, I think the world at large has read way too much into this. I don't know of anyone who has tried to reproduce the results. – Steve Nov 9 '11 at 0:09
evidently Derwent is trying to reproduce the results! They are all over the media for this because Twitter is trendy. Not only is it free advertising, it may end-run around rules about hedge fund advertising. While I agree with you that the results are very unlikely to be reproducible, it doesn't much matter if the hype gets them clients. – shabbychef Nov 9 '11 at 5:09
The Bollen paper is very dubious; it has been thoroughly discredited here: sellthenews.tumblr.com/post/21067996377/noitdoesnot – shabbychef Jun 30 '12 at 5:42

The was an article about how ten years ago some option market-makers were crawling online discussions forums (like the Yahoo! stock board) to see what people talked about.

If discussion about some ticker spiked, they assumed the volatility for that ticker will increase and factored this into their pricing models. Of course, most of the time is was because of news related to that stock, but not always (think about pump'n'dump schemes).

I assume that today they are much more sophisticated about this.

share|improve this answer
10 years ... that's interesting. It'd be great if you could dig out a link. – Ryogi Nov 3 '11 at 23:10
I've found this paper from 2003 freepatentsonline.com/y2003/0135445.html – siamii Nov 8 '11 at 21:44

A related idea is to text mine news and use it for predicting movements of stock market. See this NYTimes article for more information.

share|improve this answer

There is a product called Wall Street Birds which can be used for trading. Still in invitation mode. Another sentiment based product is Piqqem

share|improve this answer
Wolframalpha and freebase also can be used for this purpose. Addendum to the above. Also see Google Refine. – Suminda Sirinath S. Dharmasena Feb 10 '12 at 15:48
Wall Street Birds seems to have a .NET API. Interesting. – Samik R Apr 10 '12 at 21:04

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.