Lydia Manikonda

Arizona State University

Crowdsourced Planning

Crowdsourcing is gaining attention from different research communities to deal with issues that are difficult for machines to handle but are trivial for humans. Can we involve automated planners in this process to improve the efficiency of the task that humans are dealing with?

In the past, mixed initiative planning involves automated planners taking help from humans to perform the planning. But in this project, automated planners come to help the humans perform planning efficiently. The domain of interest currently is Tour planning where the goal is to generate efficient travel plans. This process of involving planners can bring challenges like interpreting the suggestions provided by the crowd and steering them. When the crowd is developing tour plans, automated planner can be used to steer the humans in an effective manner in order to generate high quality tour plans. This project involves developing the entire framework of crowdsourcing in collaboration with planners and also by addressing the challenges to perform plan recognition.

Tour Planner

Instagram Content Analysis

We are conducting the quantitative and qualitative analysis of Instagram. Essentially, we focus on the kinds of pictures posted on Instagram and the types of users existing on this platform. We also try to understand the correlation between the different categories of pictures posted on this platform. We use computer vision techniques to examine the photo content. Our results reveal several insights about Instagram which were never studies before, that include:

  • Eight popular photo categories
  • Five distinct types of Instagram users in terms of their posted photos
  • A user's audience (number of followers) is independent of his/her shared photos on Instagram

Understanding the usage of multiple Online Social Networks (OSNs) is of significant research interest as it helps in identifying the distinguishing traits of each social media platform that contribute to its continued existence. A comparison between two OSNs is particularly useful when it is done on the representative set of users holding active accounts on both the platforms. In this research, we collected a set of users holding accounts on both Twitter and Instagram. An extensive textual and visual analysis on the media content posted by these users reveals that these platforms are indeed perceived differently at a fundamental level with Instagram engaging more of the users’ heart and Twitter capturing more of their mind. These differences get reflected in the linguistic, topical and visual aspects of the user posts.


Language Dynamics of Users of Online Weight-loss Communities

Is there any correlation between the way users on weight-loss communities talk and their patterns of losing weight?

Our research focuses on answering this question interms of the way users on the online weight-loss communities ask questions, the kind of sentiment involved in their talk and how to detect lurkers on these communities. Essentially, we study the language of users in correlation with their weight dynamics, where each user check-ins his/her weight every week. Online communities bring in many challenges in terms of processing the text as there can be lots of irrelevant information while dealing with ambiguity of the textual references.

Weight Loss