Gaurav Paruthi

I am a PhD candidate in iSchool at University of Michigan, advised by Mark Newman.
My work centers around context-aware computing. Consider my CV and Projects for more info.
Contact me:   · · · ·

Projects

  • Crowd-generated Contextualized Behavior Change Messages

    HCI, UbiComp, Health

    We present a crowdsourcing algorithm that enables paid workers to create context-appropriate and tone-sensitive behavior messages. Our approach is the first to leverage crowd-workers for generating “fresh” messages.

  • Recommending Lifestyle and Behavior Changes -

    HCI, UbiComp, Health

    Modeling people’s context, habits, lifestyles, wellness goals and their achievements and failures allow interventions that are personalized to the needs of the individual. Proposed a recommender system that leverages user-similariy to recommend the most effective lifestyle choices.

  • Interactive DVD as a Platform for Education

    HCI, ICTD, Education

    We present a crowdsourcing algorithm that enables paid workers to create context-appropriate and tone-sensitive behavior messages. Our approach is the first to leverage crowd-workers for generating “fresh” messages.

  • StoneSoup: A Community Driven Proactive Display

    HCI, UbiComp, CSCW

    A community driven proactive display system that detects the nearby users and allow its users to personalize content for the display with reduced the burden of updating content.

  • Understanding Lending Behaviors on Online Microlending Platforms: The Case for Kiva

    Data Mining, ICTD, Micro-finance

    Our in-depth analysis of a microlending website shows that lenders appear to lend more to highly rated institutions, and with what appears to be better planned lending decisions; and that smaller, homogeneous teams seem to drive more lending activity and to achieve larger team lending agreements.

  • Twitter Quakes -

    DataMining

    Developed a technique to measure the impact of global events on different geographic locations. Our technique measures significant increase in the volume of tweets during specific time periods, which are generally caused by some significant happening of the event.


News