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.

The internet contains billions of documents. We show that document frequencies in large decentralized textual databases can capture the cross-sectional variation in the occurrence frequencies of social phenomena. We characterize the econometric conditions under which such proxying is likely. We also propose using recently-introduced internet search volume indexes as proxies for fundamental locational traits, and discuss their advantages and limitations. We then successfully proxy for a number of economic and demographic variables in US cities and states. We further obtain document-frequency measures of corruption by country and US state and replicate the econometric results of previous research studying its covariates. Finally, we provide the first measure of corruption in American cities. Poverty, population size, service-sector orientation, and ethnic fragmentation are shown to predict higher levels of corruption in urban America.

See next page for additional authors Follow this and additional works at:

Abstract: During the last decade the Internet has become an increasingly important source for gathering company related information. We employ Wikipedia editing frequency as an instrument that captures the degree in which the population is engaged with the processing of company-related information. We find that firms whose information is processed by the population more frequently are associated with lower analysts' forecast errors, smaller analysts' forecast dispersions, and significant changes in bid-ask spreads on analysts' recommendation days. These results indicate that information processing over the Internet is related to the degree to which investors and analysts are informed about companies. JEL Classification: G14 G24

Abstract—An ocean of data is available on the web. From this ocean of data, information can in theory be extracted and used by analysts for detecting emergent trends (trend spotting). However, to do this manually is a daunting and nearly impossible task. We describe a semi-automatic system in which data is automatically collected from selected sources, and to which linguistic analysis is applied to extract e.g., entities and events. After combining the extracted information with human intelligence reports, the results are visualized to the user of the system who can interact with it in order to obtain a better awareness of historic as well as emergent trends. A prototype of the proposed system has been implemented and some initial results are presented in the paper. Index Terms—Information fusion; trend analysis; trend spot-ting; web harvesting; web mining I.

ii Using consumer satisfaction as an example of complex communication and a virtual world as a mediating platform, a novel framework for eliciting consumer satisfaction has been developed. Consumer satisfaction is a key element for business success, while the elicitation of satisfaction perceptions from consumers can help vendors to assess and to improve their business performance. The objectives here are: how consumer satisfaction is defined, understood and measured; how virtual worlds function, both as a platform and a product; how users typically perceive their experiences in virtual worlds; and how consumer satisfaction metrics can be translated into a virtual environment. Second Life is used as an enabling technology for gathering requirements as well as for the construction, refinement and validation of the framework. Second Life is a virtual world, a multi-user, 3D, immersive environment, which has its own internal economy. The choice of using this social virtual world was due to Second Life being a resilient and widely used platform.

Manipulation of social media affects perceptions of candidates and compromises decision-making. C

The increasingly widespread use of the internet by both businesses and consumers has led to the creation of a potentially useful data source: information on internet search behaviour. Search engine providers keep a record of the searches entered on their website. Some of this information has been made publicly available, enabling users to track the popularity of an extensive range of search terms. This vast database could be used to analyse various issues. For example data on searches for ‘flatscreen televisions ’ and ‘fridges ’ could help to analyse how demand for durable goods has changed over time. Internet search data have the potential to be useful for economic policy making. Monitoring current economic activity closely is an important aspect of policymaking, but official economic statistics are generally published with a lag.

Our emotional state influences our choices. Research on how it happens usually comes from the lab. We know rela tively little about how real world emotions affect real world settings, like financial markets. Here, we demonstrate that estimating emotions from weblogs provides novel informa tion about future stock market prices. That is, it provides information not already apparent from market data. Specifi cally, 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 expres sions 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 pri marily discusses daily life, can anticipate changes in a seem ingly unrelated system. Beyond this, the results suggest new ways to gauge public opinion and predict its impact.

Online product buzz refers to an online expression of interest in a product, such as online product reviews, blog posts and search trends. We study how well online buzz predicts actual sales across different phases in the product lifecycle. Using data from smartphone sales of a Dutch online retailer, we demonstrate a 28 % overall increase in forecasting accuracy when measures of online product buzz variables are take into account. The value of online product buzz shows especially in early sales forecasting, an area in which traditional forecasting models have substantial difficulties. In early sales, local search trends, subscriptions for stock notifications and pageviews were important predictors and forecasts were on average improved by 44%. For mature sales, accuracy improved by 10%, with the most important predictors being on- and offsite reviews and, again, pageviews. These results also suggest different drivers of sales across phases in the product