Mobile Computing, Applications, and Services. Third International Conference, MobiCASE 2011, Los Angeles, CA, USA, October 24-27, 2011. Revised Selected Papers

Research Article

Rankr: A Mobile System for Crowdsourcing Opinions

Download215 downloads
  • @INPROCEEDINGS{10.1007/978-3-642-32320-1_2,
        author={Yarun Luon and Christina Aperjis and Bernardo Huberman},
        title={Rankr: A Mobile System for Crowdsourcing Opinions},
        proceedings={Mobile Computing, Applications, and Services. Third International Conference, MobiCASE 2011, Los Angeles, CA, USA, October 24-27, 2011. Revised Selected Papers},
        proceedings_a={MOBICASE},
        year={2012},
        month={10},
        keywords={ranking mobility crowdsourcing incentives user interfaces},
        doi={10.1007/978-3-642-32320-1_2}
    }
    
  • Yarun Luon
    Christina Aperjis
    Bernardo Huberman
    Year: 2012
    Rankr: A Mobile System for Crowdsourcing Opinions
    MOBICASE
    Springer
    DOI: 10.1007/978-3-642-32320-1_2
Yarun Luon1,*, Christina Aperjis1,*, Bernardo Huberman1,*
  • 1: HP Labs
*Contact email: yarun.luon@hp.com, christina.aperjis@hp.com, bernardo.huberman@hp.com

Abstract

Evaluating large sets of items, such as business ideas, is a difficult task. While no one person has time to evaluate all the items, many people can contribute by each evaluating a few. Moreover, given the mobility of people, it is useful to allow them to evaluate items from their mobile devices. We present the design and implementation of a mobile service, , which provides a lightweight and efficient way to crowdsource the relative ranking of ideas, photos, or priorities through a series of pairwise comparisons. We discover that users prefer viewing two items simultaneously versus viewing one image at a time with better fidelity. Additionally, we developed an algorithm that determines the next most useful pair of candidates a user can evaluate to maximize the information gained while minimizing the number of votes required. Voters do not need to compare and manually rank all of the candidates.