Research

Research projects at iiD focus on building connections. We encourage crosspollination of ideas across disciplines, and to develop new forms of collaboration that will advance research and education across the full spectrum of disciplines at Duke. The topics below show areas of research focus at iiD. See all of our research.

A Durham-based startup, founded by Duke professors Larry Carin (ECE) and Ricardo Henao (B&B), is looking for interns excited to develop skills and gain experience in data science and machine learning. This is an opportunity to learn from and contribute to a new and growing company focused on cutting-edge machine learning and deep learning technology.

The aim of our data expeditions course was to give students in Bio 190S-0.2, a summer session course in sensory systems, an introduction to how real data may actually look and how they may actually be analyzed. Over the course of a two-hour class session, 16 students ranging from 16-22 years old were given the opportunity to explore a dataset on the color vision capabilities of three species of cleaner shrimp.

Matt and Ken led two labs for the engineering section of STA 111/130, an introductory course in statistics and probability. The lab assignments were written by Matt and Ken in order to bridge the gap between introductory linear regression, which is often explained in terms of a static, complete dataset, and time series analysis, which is not a common topic in introductory courses. 

This paper addresses analysis of heterogeneous data, such as ordered, categorical, real and count data. Such data are of interest in our motivating application, cognitive and brain science, in which subjects may answer questionnaires, and also (separately) undergo fMRI interrogation. A contribution of this paper concerns the joint analysis of how people answer questionnaires and how their brain responds to external stimuli (here visual), the latter measured via fMRI.

Devri Adams (Environmental Science), Annie Lott (Statistics), and Camila Vargas Restrepo (Visual Media Studies, Psychology) spent ten weeks creating interactive and exploratory visualizations of ecological data. They worked with over sixty years of data collected at the Hubbard Brook Experimental Forest (HBEF) in New Hampshire.

Ana Galvez (Cultural and Evolutionary Anthropology), Xinyu Li (Biology), and Jonathan Rub (Math, Computer Science) spent ten weeks studying the impact of diet on organ and bone growth in developing laboratory rats. The goal was to provide insight into the growth dynamics of these model organisms that could eventually be generalized to inform research on human development.

Robbie Ha (Computer Science, Statistics), Peilin Lai  (Computer Science, Mathematics), and Alejandro Ortega (Mathematics) spent ten weeks analyzing the content and dissemination of images of the Syrian refugee crisis, as part of a general data-driven investigation of Western photojournalism and how it has contributed to our understanding of this crisis.

Over ten weeks, Computer Science majors Daniel Bass-Blue and Susie Choi joined forces with Biomedical Engineering major Ellie Wood to prototype interactive interfaces from Type II diabetics' mobile health data. Their specific goals were to encourage patient self-management and to effectively inform clinicians about patient behavior between visits.

Over ten weeks, Computer Science Majors Amber Strange and Jackson Dellinger joined forces with Psychology major Rachel Buchanan to perform a data-driven analysis of mental health intervention practices by Durham Police Department. They worked closely with leadership from the Durham Crisis Intervention Team (CIT) Collaborative, made up of officers who have completed 40 hours of specialized training in mental illness and crisis intervention techniques.

Building off the work of a 2016 Data+ teamYu Chen (Economics), Peter Hase (Statistics), and Ziwei Zhao (Mathematics), spent ten weeks working closely with analytical leadership at Duke's Office of University Development. The project goal was to identify distinguishing characteristics of major alumni donors and to model their lifetime giving behavior.

A team of students led by Duke mathematician Marc Ryser and University of Southern California Pathology professor Darryl Shibata will characterize phenotypic evolution during the growth of human colorectal tumors. 

Graduate Students: Kendra Kaiser and John Mallard

Faculty: Michael O’Driscoll

Course: Landscape Hydrology, EOS 323/723

A team of students led by Dr. Shanna Sprinkle of Duke Surgery will combine success metrics of Duke Surgery residents from a set of databases and create a user interface for residency program directors and possibly residents themselves to view and better understand residency program performance.

Lauren Fox (Cultural Anthropology) and Elizabeth Ratliff (Statistics, Global Health) spent ten weeks analyzing and mapping pedestrian, bicycle, and motor vehicle data provided by Durham's Department of Transportation. This project was a continuation of a seminar on "ghost bikes" taught by Prof. Harris Solomon.

Boning Li (Masters Electrical and Computer Engineering), Ben Brigman (Electrical and Computer Engineering), Gouttham Chandrasekar (Electrical and Computer Engineering), Shamikh Hossain (Computer Science, Economics), and Trishul Nagenalli (Electrical and Computer Engineering, Computer Science) spent ten weeks creating datasets of electricity access indicators that can be used to train a classifier to detect electrified villages. This coming academic year, a Bass Connections Team will use these datasets to automatically find power plants and map electricity infrastructure.

Felicia Chen (Computer Science, Statistics), Nikkhil Pulimood (Computer Science, Mathematics), and James Wang (Statistics, Public Policy) spent ten weeks working with Counter Tools, a local nonprofit that provides support to over a dozen state health departments. The project goal was to understand how open source data can lead to the creation of a national database of tobacco retailers.

Selen Berkman (ECE, CompSci), Sammy Garland (Math), and Aaron VanSteinberg (CompSci, English) spent ten weeks undertaking a data-driven analysis of the representation of women in film and in the film industry, with special attention to a metric called the Bechdel Test. They worked with data from a number of sources, including fivethirtyeight.com and the-numbers.com.

Over ten weeks, BME and ECE majors Serge Assaad and Mark Chen joined forces with Mechanical Engineering Masters student Guangshen Ma to automate the diagnosis of vascular anomalies from Doppler Ultrasound data, with goals of improving diagnostic accuracy and reducing physician time spent on simple diagnoses. They worked closely with Duke Surgeon Dr. Leila Mureebe and Civil and Environmental Engineering Professor Wilkins Aquino.

Over ten weeks, Math/CompSci majors Benjamin Chesnut and Frederick Xu joined forces with International Comparative Studies major Katharyn Loweth to understand the myriad academic pathways traveled by undergraduate students at Duke. They focused on data from Mathematics and the Duke Global Health Institute, and worked closely with departmental leadership from both areas.

Liuyi Zhu (Computer Science, Math), Gilad Amitai (Masters, Statistics), Raphael Kim (Computer Science, Mechanical Engineering), and Andreas Badea (East Chapel Hill High School) spent ten weeks streamlining and automating the process of electronically rejuvenating medieval artwork. They used a 14th-century altarpiece by Francescussio Ghissi as a working example.

Angelo Bonomi (Chemistry), Remy Kassem (ECE, Math), and Han (Alessandra) Zhang (Biology, CompSci) spent ten weeks analyzing data from social networks for communities of people facing chronic conditions. The social network data, provided by MyHealth Teams, contained information shared by community members about their diagnoses, symptoms, co-morbidities, treatments, and details about each treatment.

John Benhart (CompSci, Math) and Esko Brummel (Masters in Bioethics and Science Policy) spent ten weeks analyzing current and potential scholarly collaborations within the community of Duke faculty. They worked closely with the leadership of the Scholars@Duke database.

Zijing Huang (Statistics, Finance), Artem Streltsov (Masters Economics), and Frank Yin (ECE, CompSci, Math) spent ten weeks exploring how Internet of Things (IoT) data could be used to understand potential online financial behavior. They worked closely with analytical and strategic personnel from TD Bank, who provided them with a massive dataset compiled by Epsilon, a global company that specializes in data-driven marketing.

Furthering the work of a 2016 Data+ team in predictive modeling of pancreatic cancer from electronic medical record (EMR) data, students Siwei Zhang (Masters Biostatistics) and Jake Ukleja (Computer Science) spent ten weeks building a model to predict pancreatic cancer from Electronic Medical Records (EMR) data. They worked with nine years worth of EMR data, including ICD9 diagnostic codes, that contained records from over 200,000 patients.

Over ten weeks, Mathematics/Economics majors Khuong (Lucas) Do and Jason Law joined forces with Analytical Political Economy Masters student Feixiao Chen to analyze the spati-temporal distribution of birth addresses in North Carolina. The goal of the project was to understand how/whether the distributions of different demographic categories (white/black, married/unmarried, etc.) differed, and how these differences connected to a variety of socioeconomic indicators.

William Willis (Mechanical Engineering, Physics) and Qitong Gao (Masters Mechanical Engineering) spent ten weeks with the goal of mapping the ocean floor autonomously with high resolution and high efficiency. Their efforts were part of a team taking part in the Shell Ocean Discovery XPRIZE, and they made extensive use of simulation software built from Bellhop, an open-source program distributed by HLS Research.

Over ten weeks, Public Policy major Amy Jiang and Mathematics and Computer Science major Kelly Zhang joined forces with Economics Masters student Amirhossein Khoshro to investigate academic hiring patterns across American universities, as well as analyzing the educational background of faculty. They worked closely with Academic Analytics, a provider of data and solutions for universities in the U.S. and the U.K.

Linda Adams (CompSci), Amanda Jankowski (Sociology, Global Health), and Jessica Needleman (Statistics/Economics) spent ten weeks prototyping small-area mapping of public-health information within the Durham Neighborhood Compass, with a focus on mortality data. They worked closely with the director of DataWorks NC, an independent data intermediary dedicated to democratizing the use of quantitative information.

Gary Koplik (Masters in Economics and Computation) and Matt Tribby (CompSci, Statistics) spent ten weeks investigating the burden of rare diseases on the Duke University Health System (DUHS). They worked with a massive set of ICD diagnosis codes and visit data provided by DUHS.

Over ten weeks, Biology major Jacob Sumner and Neuroscience major Julianna Zhang joined forces with Biostatistics Masters student Jing Lyu to analyze potential drug diversion in the Duke Medical Center. Early detection of drug diversion assists health care providers in helping patients recover from their condition, as well as mitigate the effects on any patients under their care.

Graduate Student: Jacob Coleman, 3rd year Ph.D. student in Statistical Science

Faculty Instructor: Colin Rundel

Class: STA 112, Data Science

Students in the Performance and Technology Class create a series of performances that explore the interface between society and our machines. With the theme of the cloud to guide them, they have created increasingly complex art using digital media, microcontrollers, and motion tracking. Their work will be on display at the Duke Choreolab 2016.

Graduate student: Hamza Ghadyali          

Faculty instructor: Dr. Paul Bendich

Course: MATH 412 – Topology with Applications

With the significant international consequences of recent outbreaks, the ITP Lab conducted extensive stakeholder interviews and macro-level health policy analysis to expose gaps in pandemic preparedness and develop legal frameworks for future threats. 

This project summarizes the existing sample agreements from different institutions, analyzes the key contractual issues in the formation of alliances, and develops master charts of legal provisions to compare different approaches, to provide a reference for the formation of new alliances in the era of epidemic disease outbreaks. 

A virtual reality system to recreate the archaeological experience using data and 3D models from the neolithic site of Çatalhöyük, in Anatolia, Turkey. 

How well and in what ways do governments communicate with their citizens? How do governments analyze data and create visualizations to promote public access to government information? 

With the significant international consequences of recent outbreaks, the ITP Lab conducted extensive stakeholder interviews and macro-level health policy analysis to expose gaps in pandemic preparedness and develop legal frameworks for future threats. 

Paclitaxel (Taxol) is a small molecule drug belonging to the taxane family. It is one of the most commonly used chemotherapeutics, used for treatment of many cancers, as a monotherapy or in combination with other drugs to treat breast, lung and ovarian cancer as well as Kaposi’s sarcoma. Taxol is on the World Health Organization’s (WHO) List of Essential Medicines, a list that includes most the important medications for basic health. The worldwide demand for paclitaxel is exceeding the current supply. 

This project transforms an inaccessible audio archive of historic North Carolina folk music colllected by Frank Clyde Brown in the 1920s-40s into a vital, publicly accessible digital archive and museum exhibition. 

Imagine a world where we understand how to detect mental health and developmental problems in early childhood so that we can intervene early in life and prevent future suffering and impairment. This is a challenge that can only be addressed by an interdisciplinary team of computational people with child psychiatrists and neuroscientists who can integrate and mine knowledge from cross-cultural and global data.

Lineage Logistics is the second largest cold storage network in the world, playing a critical role in multiple global supply chains. We store and transport temperature-sensitive commodities (about 30 billion lbs per year) in a large network of warehouses, trucks and rail cars. Our inventories include everything from Boeing’s carbon fiber to your 4th of July baby-back ribs.

Nonnegative matrix factorization (NMF) has an established reputation as a useful data analysis technique in numerous applications. However, its usage in practical situations is undergoing challenges in recent years.The fundamental factor to this is the increasingly growing size of the datasets available and needed in the information sciences. To address this, in this work we propose to use structured random compression, that is, random projections that exploit the data structure, for two NMF variants: classical and separable. In separable NMF (SNMF) the left factors are a subset of the columns of the input matrix. We present suitable formulations for each problem, dealing with different representative algorithms within each one.

In this work, we turn musical audio time series data into shapes for various tasks in music matching and musical structure understanding. 

Large-scale databases from the social, behavioral, and economic sciences offer enormous potential benefits to society. However, as most stewards of social science data are acutely aware, wide-scale dissemination of such data can result in unintended disclosures of data subjects' identities and sensitive attributes, thereby violating promises–and in some instances laws to protect data subjects' privacy and confidentiality. 

The Triangle Census Research Network (TCRN) is an interdisciplinary team of researchers from Duke University and the National Institute of Statistical Sciences dedicated to improving the way that federal statistical agencies collect, analyze, and disseminate data to the public.

We present a framework for high-dimensional regression using the GMRA data structure. In analogy to a classical wavelet decomposition of function spaces, a GMRA is a tree-based decomposition of a data set into local linear projections.

In this Data Expedition, Duke undergraduates were introduced to a real world traffic citation data set. Provided by Dr. Frank R. Baumgartner, a political scientist at UNC, the data consist of 15 years of traffic stops, with over 18 million observations of 53 variables.

Dr. Guillermo Sapiro, professor in Pratt School of Engineering at Duke University, conducts ongoing autism research. Using image processing, he attempts to program a computer to detect whether babies (around eight to 14 months of age) display a sign of autism. This very early detection enables doctors to train these babies (when their brain plasticity is high) to behave in ways to counter the behavioral limitations autism imposes, thus allowing these babies to act more normally as they grow up. 

Using social network analysis to predict survival in large-brained mammals.

Graduate students: Aaron Berdanier and Matt Kwit, University Program in Ecology & Nicholas School of the Environment

This data expedition introduced students to “sliding windows and persistence” on time series data, which is an algorithm to turn one dimensional time series into a geometric curve in high dimensions, and to quantitatively analyze hybrid geometric/topological properties of the resulting curve such as “loopiness” and “wiggliness.”

In this project, we aim to solve the compressive sensing (CS) hyperspectral / video image reconstruction problem. The propose algorithm is robust to different initializations. This is useful for CS reconstruction problems where the suitable training datasets are not available.

Students learned to visualize high-dimensional gene expression data; understand genetic differences in the context of gene networks; connect genetic differences to physiological outcomes; and perform simple analyses using the R programming language.

Introduce NBA and MLB datasets to undergraduates to help them gain expertise in exploratory data analysis, data visualization, statistical inference, and predictive modeling.

Questions asked: Do males and females scent mark equally? Do lemurs scent mark equally in breeding and non-breeding seasons?

STEM education often presents a very sanitized version of the scientific enterprise. To some extent, this is necessary, but overemphasizing neat-and-tidy results and scripted protocol assignments poses the risk of failing to adequately prepare students for the real-world mess of transforming experimental data into meaningful results. The fundamental aim of this project was to guide students in processing large real-world datasets far beyond their academic comfort zone so as to give them a more realistic understanding of how science works.

What drove the prices for paintings in 18th Century Paris?

A new model is developed for joint analysis of ordered, categorical, real and count data. In the motivating application, the ordered and categorical data are answers to questionnaires, the (word) count data correspond to the text questions from the questionnaires, and the real data correspond to fMRI responses for each subject. We also combine the analysis of these data with single-nucleotide polymorphism (SNP) data from each individual. 

The sub-thalamic nucleus (STN) within the sub-cortical region of the Basal ganglia is a crucial targeting structure for Deep brain stimulation (DBS) surgery, in particular for alleviating Parkinson’s disease (PD) symptoms. Volumetric segmentation of such small and complex structure, which is elusive in clinical MRI protocols, is thereby a pre-requisite process for reliable DBS targeting. While direct visualization and localization of the STN is facilitated with advanced high-field 7T MR imaging, such high fields are not always clinically available. 

Volumetric segmentation of sub-cortical structures such as the basal ganglia and thalamus is necessary for non-invasive diagnosis and neurosurgery planning. This is a challenging problem due in part to limited boundary information between structures, similar intensity profiles across the different structures, and low contrast data.

Intelligent mobile sensor agent can adapt to heterogeneous environmental conditions, to achieve the optimal performance, such as demining, maneuvering target tracking. 

Successful high-resolution signal reconstruction -- in problems ranging from astronomy to biology to medical imaging -- depends crucially our ability to make the most out of indirect, incomplete, a