Research

The projects documented here could not have been possible without folks I collaborated with at different stages. Do checkout their info in the collaborators section.

On-going projects:

On going in the form of data collection, data analysis, software development, and/or maintenance.



TIME: Temporal Influences in Movement and Exercise

pain_scknessTIME study is conducted to examine different micro-temporal processes that influence health behavior change and maintenance. We developed a smartphone and a smartwatch (called TIME app) to capture variations in different health constructs using passive sensors, app usage, mobile-based ecological momentary assessment, and microinteractions-based ecological momentary assessment. Read more …


Signaligner Pro: Algorithm-assisted multi day sensor data annotation

Screenshot 2020-03-13 at 1.28.35 AM

Signaligner-Pro is an interactive tool for algorithm-assisted exploration and annotation of raw accelerometer data. The tool can be used by researchers using raw accelerometer data to support research in activity recognition/machine learning, exercise science, and sleep quality research among others. In particular, the tool can be used to visualize, explore, and annotate multi-day high sampling rate raw sensor data. In addition, it provides opportunities to researchers to collaborate when annotating raw data. Read more …


μEMA: Microinteractions-based ecological momentary assessment

sitting-right_nowμEMA (micro-EMA) is microinteractions-based ecological momentary assessment (EMA) tool for digital phenotyping where all the prompts are single questions with a “Yes/No” type answer. These self-report prompts can be responded to using quick glanceable microinteractions that take ~2s (like checking the time on the watch). With this cognitive simplicity, μEMA could gather self-reported behavioral data with much higher temporal density than traditional EMA. Read more …


Mix-WILD: A freeware program for mixed model analysis with intensive longitudinal data

MixWILD (Mixed model analysis with Intensive Longitudinal Data) is a desktop GUI-based application for examining the effects of variance and slope of time-varying variables in intensive longitudinal data, especially the ones collected using ecological momentary assessments. Instead of using programming interfaces such as SAS NLMIXED, Python or R, MixWILD allows researchers to use a simple GUI to build complex multilevel models. Read more …



Past projects:

Projects completed with no more software development, data collection, and/or analysis.



Human-computation game for sensor data annotation

In this project, we explored human-computation games to crowdsource annotations on raw accelerometer data from casual game players. We designed two games Mobots and Signaligner and evaluated them using Amazon Mechanical Turk. The purpose was to enable activity recognition researchers to generate training datasets from the real-world acceleration data with the help of casual game players.  Read more …


Sideline: An in situ concussion assessment tool

Sideline allows parents, coaches of informal sports clubs, and players to assess the possibility of a concussion on the sports field. It is implemented on the Android platform using clinically approved symptom assessments, working memory, visual memory, reaction time, and physical balance tests. Read more …


Finding Astro: Mobile exergame to promote physical activity

Finding Astro is an Android-based experimental exergame intended to promote small bouts of physical activity among young children. The player finds a missing astronaut in space by moving the smartphone in 3-D space. The direction of the motion is directed by specific exercise moves designed to induce exertion. Read more …


Exploring social capital in low-SES communities using mobile technology

We conducted a qualitative study of a community program that seeks to empower low-income neighborhood residents through a mobile application that connects them to local resources. Our findings highlight how the application and offline socio-organizational mechanisms worked in tandem to create gateways for capital building-sparking connections (to people and opportunities) that residents leveraged with varying motivations and outcomes. Read more …


User experience (UX) research at Samsung Electronics

I worked on a variety of projects at Samsung electronics in the areas of personal health, text input, media-multitasking, and alternative OS UIs for low cost smartphones in re-emerging markets. Here, I discuss briefly some of the work done during my stint there. Read more …


Digital chameleons: Investigating the effect of mimicry by a virtual agent

As part of my research internship at the Eindhoven University of Technology (TU/e), The Netherlands, I carried out a between-subject experiment to study the effect of mimicry by a virtual agent on user’s perceived trustworthiness. This study establishes that social mimicry (like mimicking the head movements) results in higher trust on the agent, which is mediated through the liking for that agent. Read more …