Widespread Worry and the Stock Market

Saturday, March 13th, 2010

UPDATE: I have released the classifiers, R scripts and aggregate data from this paper. The archive has a README to get you started and some example Java showing how to use the classifiers. Get it here.

I have a new paper at ICWSM 2010. I’m really looking forward to all the great work in the program. The central thesis of my paper: estimating anxiety, worry and fear from blogs provides some novel information about future stock market prices.

ABSTRACT: Our emotional state influences our choices. Research on how it happens usually comes from the lab. We know relatively little about how real world emotions affect real world settings, like financial markets. Here, we demonstrate that estimating emotions from weblogs provides novel information about future stock market prices. That is, it provides information not already apparent from market data. Specifically, we estimate anxiety, worry and fear from a dataset of over 20 million posts made on the site LiveJournal. Using a Granger-causal framework, we find that increases in expressions of anxiety, evidenced by computationally-identified linguistic features, predict downward pressure on the S&P 500 index. We also present a confirmation of this result via Monte Carlo simulation. The findings show how the mood of millions in a large online community, even one that primarily discusses daily life, can anticipate changes in a seemingly unrelated system. Beyond this, the results suggest new ways to gauge public opinion and predict its impact.

pdf Widespread Worry and the Stock Market.
Proc. ICWSM, 2010.

CHI 2009: Predicting Tie Strength">CHI 2009: Predicting Tie Strength

Wednesday, January 14th, 2009

Social media treats all users the same: trusted friend or total stranger, with little or nothing in between. In reality, relationships fall everywhere along this spectrum, a topic social science has investigated for decades under the theme of tie strength. Our work bridges this gap between theory and practice. In this paper, we present a predictive model that maps social media data to tie strength. The model builds on a dataset of over 2,000 social media ties and performs quite well, distinguishing between strong and weak ties with over 85% accuracy. We complement these quantitative findings with interviews that unpack the relationships we could not predict. The paper concludes by illustrating how modeling tie strength can improve social media design elements, including privacy controls, message routing, friend introductions and information prioritization.

We won best paper!

pdf Predicting Tie Strength With Social Media.
Proc. CHI, 2009.

HICSS 2009: Blogs Are Echo Chambers">HICSS 2009: Blogs Are Echo Chambers

Wednesday, October 1st, 2008

> 3:1 agreement

Tony, a co-author of this work, dreamt up the very clever title (see full citation at end of post). I particularly love the use of the highly academic colon. I will present it at the Social Spaces minitrack, part of the Digital Media track (all very hierarchical). Soon I will release the data, code and algorithm specifics from this paper. I included urls in the text of the paper, so I really need to post it soon. I was very happy to see this work come together, and I very much look forward to seeing some of the other work at the minitrack. Plus, Hawaii in January (+ baby depending on how fussy she seems near ticket-buying time) will be awfully nice. I need to start shopping for parasols and shark repellent.

pdf Blogs Are Echo Chambers: Blogs Are Echo Chambers.
Proc. HICSS, 2009.