Who we are

Profile picture of Aritra Dasgupta
Aritra Dasgupta
Assistant Professor
Department of Data Science
Profile Picture of Jun Yuan
Jun Yuan
PhD student
Department of Informatics
Profile Picture of Vrushali Koli
Vrushali Koli
PhD student
Department of Computer Science
Kaustav Bhattacharjee
PhD student
Department of Informatics
Akm Islam
University Lecturer
Department of Data Science

Research Philosophy

We are a group of data toolsmiths who develop and study the role of visualization techniques as a transparent lens between what is computed from data and what is communicated to the human mind.

As George Bernard Shaw so eloquently said: “The single biggest problem in communication is the illusion that it has taken place.”  This is oh-so-evident in today’s age, where, information, if communicated properly, can cure diseases and fuel discoveries, but, if miscommunicated, can lead to an “infodemic” in the worst-case scenario.

To solve this conundrum, at NiiV, we pursue intelligibility as the foundational principle for making information more accessible, meaningful, and actionable to experts (e.g., doctors, climate scientists) and non-experts alike.

We operationalize this principle by visualizing data, big or small, with the ultimate goal of letting human observers see, understand, and trust the information that is often generated by black-box algorithms.

By embracing a human-centered data science approach that ultimately culminates in interactive visual analytic interfaces, we preserve the best of both worlds: the power of computational methods and that of human judgment and reasoning.

Research Highlights

Visual Analytics for Content Moderation
IEEEVIS ’21 (Short paper)
highlight 1
Conceptualizing Visual Analytic Interventions for Content Moderation, by S. Vaidya, J. Cai, S. Basu, A. Naderi, D.Y. Wohn and A. Dasgupta, IEEE VIS, 2021
MyriadCues: A Multiway Visual Comparison Tool
TVCG ’20
highlight 1
Visualization and Machine Learning
IEEEVIS ’21 TREX Workshop
highlight 1
Beyond Visual Analytics: Human-Machine Teaming for AI-Driven Data Sensemaking, by J. Wenskovitch, C. Fallon and A. Dasgupta, IEEE VIS, 2021
Privacy meets Visualization
CGF ‘20
highlight 1

Lab updates

Oct ’22: Bhattacharjee presented paper at VizSec

July ’22: Bhattacharjee interned at PNNL for summer and fall '22

Apr ’22: Yuan presented position paper at HCDS@CHI

Oct ’21: 2 papers presented at IEEEVIS: short paper track (Vaidya) and TREX workshop(Dasgupta)

Oct ’21: Dasgupta chairs session on VIS+AI at IEEEVIS

July ’21: Dasgupta receives NJIT Seed grant for research on communicative visualization

May ’21: Yuan interned at Accern for summer '21

Oct ’20: 2 papers and a poster presented at IEEEVIS ‘20

May ’20: Paper at EuroVis presented by Bhattacharjee

May ’20: Dasgupta receives grant from Hearst Corporation for data visualization training

Research Areas

Interactive Visual Comparison

Let domain scientists reason about computational model behavior and help them select the most accurate models by interactively comparing multiple facets of model performance.

EuroVis14 | TVCG14 | CISE15 | InfoVis19

Studies on Visualization Effectiveness

Conduct user studies with experts from biology and climate science domains to evaluate if and how optimal visualization design can overcome potential biases due to familiarity.

TVCG16 | CHI17 | Chapter 6,Cognitive Biases Book 18 | TVCG19

Model Explainability and Trust

Provide domain experts and model developers with tools that explain the decisions of machine learning models and help them semantically validate models.

HILDA17 | TVCG17 | VAST17 | UIST18

Visualization Perception & Design Analysis

Study and survey of the visualization design space for devising classification schemes that bridge human perception with visual encodings.

TVCG15 | CGF17 | CGF18 | VisComm18

High-Dimensional Pattern Search

Provide guidance to analysts for finding patterns in high- dimensional subspaces by devising metrics that quantify salient patterns.

InfoVis10 | LDAV12 | CGF2015 | LDAV2016

Privacy-Preserving Data Visualization

Adapt visualizations to prevent disclosure of sensitive information by developing information loss metrics that can help address the trade-off between privacy gain and loss of utility due to anonymization.

InfoVis11 | CGF12 | CGF13 | EHRVis14 | VizSec19 | EuroVis20


class="wow fadeIn"

Contact us

Get in touch with us here

  • 218 Central Ave, Newark, NJ, USA - 07102

  • aritra.dasgupta@njit.edu