Florent Frédéric Garcin

I'll summarize my research here as soon as I've time. This includes our work on a) gamification of user studies on aggregation of product reviews, b) news recommendations and the implementation of a news recommender system on the newspaper swissinfo.ch, and c) swissnoise, our crowdsourcing platform to predict outcomes of future events.

Eliciting information from the crowd

swissnoise allows users to express their opinions on various topics ranging from sports, entertainments or politics. We use this platform to experiment on information aggregation from the crowd, commonly called the wisdom of the crowds.

swissnoise is an awesome tool for e-governments, organizations and companies. It allows organizations to get insight on crucial issues in order to improve their analysis and reduce bad decision making. Depending on the needs, swissnoise can run in-house as a private solution with access controls over its users or as a semi- or fully-public solution. We think that swissnoise is usefull in cases such as policy making, product development, developing and selecting marketing campaigns, sales forecasting, product launch dates, office openings, mergers/acquisitions, and many more.

Examples of questions

  • What economic growth will result if we raise the interest rates by X%?
  • Will there be a famine in country X if the harvest fails two times in a row?
  • Which holiday schedule is the best?
  • Should we build another railway track or a new highway between cities A and B?
  • Will increasing the highway tax decrease the traffic congestion?
  • Will adding another subway line decrease the popularity of subway line X?
  • Should we open a branch/office in country X?
  • How much sales will increase if we decrease the price of product X by 10%?
  • What will be the sales of product X for Q1?
  • Is product X going to launch on time?
  • What is the average connection speed of internet provider X?
  • ...

We have developed and implemented several information aggregation mechanisms such as prediction markets, simple surveys and dynamic opinion polls. swissnoise relies on gamification, in which we wrap these mechanisms into a fun and interesting game to increase users engagement.

Prediction markets allow users to update their predictions at any time. However, predictions markets need realized outcomes, i.e. one of the outcome must be observed. For instance, a typical question for prediction markets could be "Will it snow tomorrow in Geneva?" The outcomes are "yes" and "no". One of these outcomes will be observed tomorrow. However in the case of hypothetical questions where outcomes might not be observed at all such as "What economic growth will result if we raise the interest rates by 0.5%?" prediction markets cannot be used.

Simple surveys solve prediction markets' issue. Surveys aggregate opinions of people at specific time. They have the advantage of not requiring observable outcomes. Although they can give insight on hypothetical questions such as "What economic growth will result if we raise the interest rates by 0.5%?", their drawback is that they do not run over time. Users cannot update their opinions.

Dynamic opinion polls are like opinion polls but run over time. In the same way as prediction markets, they are dynamic, continuous and allow users to update their opinions. They create a constant feedback loop of updated information. They have the interesting advantages of not requiring observable outcomes, like simple surveys. They bring all the advantages of prediction markets and simple surveys into a single mechanism.