IT Service Status

2023-2024 Data Science, Statistics and Visualization Services Annual Report

During the 2023-2024 academic year, the Data Science, Statistics, and Visualization (DSSV) team within Northwestern IT Research Computing and Data Services supported the data science and visualization needs of Northwestern researchers in multiple ways:

Workshops provide opportunities to learn new skills and explore tools and analysis methods via high-quality training at no or low cost to the researchers.

Individual consultations help researchers troubleshoot issues, explore new analysis approaches, get assistance with data manipulation, interpret results, and create visualizations.
Bring Your Own Data (BYOD) working groups support researchers undertaking data-intensive work by connecting them to DSSV staff and other researchers across the university for accountability and advice.
Collaborative projects bring together DSSV staff’s specialized skills with faculty researchers from many different domains to undertake impactful research.

Workshops

The DSSV team offered virtual and in-person workshops on both the Evanston and Chicago campuses. An increase in staff size allowed us to teach more workshops than we did in previous years. We were also able to teach our popular summer boot camps on both campuses, where researchers spent multiple days learning introductory and intermediate R and Python skills.

In our Next Steps in R and Next Steps in Python series, we introduced many new advanced topics and launched our first AI workshops and our first statistical concepts workshops, which were taught outside of a specific coding language.

The four most popular workshops of FY24 were:

How to Code with ChatGPT
Statistical Modeling with Correlated Data
Python Fundamentals Bootcamp
RStudio Projects

Impact

70 percent increase in workshop hours offered
41 percent of workshop registrants were from the Feinberg School of Medicine—a steady rise since FY20, when it was 24 percent
32 percent of registrants from FY23 returned to register for one or more workshops in FY24
14 percent increase in the number of registrants attending one or more workshops

213

Data Science and Visualization Workshops Hours

1031

Individuals Registered for One or More Workshps

99.5%

Workshop Attendees Would Recommend a Workshop to their Colleagues

Role of Workshop Registrants

Consultations

We support researchers both through one-on-one research consultations and small Bring Your Own Data (BYOD) working groups that meet weekly over a quarter. FY24 saw the continued popularity of the genomics and visualization BYOD working groups and the introduction of a new social sciences BYOD working group.

Before FY24, our data scientists provided statistical support only where they were qualified and as availability allowed. In FY24, we welcomed two statisticians to the DSSV team which enabled us to increase statistical support as part of our consultation service. Reflecting the previously unmet need at the University, adding statistical support increased the number of consultations we completed and the number of researchers served.

Impact

62 percent increase in the number of statistical consultations achieved
23 percent increase in the number of researchers served in FY23
22 percent increase in the number of consultations achieved

311

Total Consultations

251

Individual Researchers Served

101

Statistical Consultations

Role of Researchers Requesting Consultations

DSSV consultants supported researchers with consultations on a wide range of topics, including:

Conducting a power analysis
Using Hugging Face transformer models to analyze text data
Containerizing an environment to work with local GPUs
Generating a survival curve from bulk RNA-Seq data
Setting up a GitHub account
Parsing JSON strings in Python
Applying statistical tests to microbial community data
Displaying multiple plots on a Shiny app
Using an API to collect bulk data
Fitting a mixed effect model to behavioral data
Knitting R markdown to HTML
Using LLMs and computer vision models to study archival materials
Learning and incorporating “for loops” in R code
Filtering very big data in R
Interpreting Python code borrowed from another researcher
Troubleshooting a custom MATLAB GUI
Presenting a large number of survey results in a paper
Merging multiple data files in R
Adding error bars to a plot
Clustering job descriptions using latent semantic analysis
Opening EEG files in MATLAB
Choosing the appropriate statistical model for a study with small sample size
Reviewing Hierarchical Linear Modeling to check for accuracy
Working with data from the American Housing Survey
Speeding up R code using parallel processing
Creating a complex heatmap
Exporting Python image files to have editable text
Working with .dat files
Selecting a model for a meta-analysis
Fitting a negative binomial model in R

Collaborative Project Support

In addition to projects in data collection, analysis, and visualization, we assisted with grant proposal preparation and exploratory investigations to determine proof of concept, allowing us to assist faculty beginning at the earliest stages of a project. We also began incorporating LLM models into project workflows as appropriate and took on the development of our first AI research tool, with more to follow in FY25.

Faculty Projects Active in FY24

Northwestern Schools Supported

Project Highlights

The team continued several collaborations from previous years and started new ones. A few projects are highlighted below. You can read more about the DSSV team’s projects in the Project Portfolio.

RainFlow

A new software product, Rainflow, was published on the Apple App Store in partnership with Luis Amaral (McCormick). Rainflow provides a user-friendly interface to analyze flow cytometry data with an emphasis on reproducibility and data validation. RCDS staff coded the core algorithms used in the app, designed the software program flow, wrote the documentation, and created simulated data for testing and validation.

EpiPred

For Gemma Carvill (Feinberg), our team made significant improvements to the EpiPred web tool that allows families and researchers to explore predicted clinical outcomes of specific genetic variants related to epilepsy.

SSO-Protected Web Server

We began work building an SSO-protected web server for Sumit Dhar's (School of Communication) lab. The server will host multiple Shiny applications to explore clinical audiology data.

AI Simulation Tool

RCDS staff collaborated with Pritzker Law faculty Esther Barron, Stephen Reed, and colleagues in IT Media and Technology Innovation to develop and launch an AI simulation tool to train law students to communicate and work with entrepreneurs.

Police Use of Force

University staff and graduate student consultants provided data analysis, visualization, and methodological input for a recently published newspaper article by Kari Lydersen (Medill) and a cohort of Medill students on police use of force in response to mental health calls in a dozen large US cities.

Research Network Diagrams

Network diagrams to demonstrate the strength and type of connections across members of a research network were created for Thomas McDade (Weinberg) to enhance a grant proposal.

Student Consultation Program

Each year, talented graduate and undergraduate students at Northwestern apply to our highly competitive Student Data Science Consultant Program. Students come into the program with experience in data science, statistics, or visualization and develop their professional and advanced technical skills through hands-on work with other Northwestern researchers.

50%

Percent of Total Consultations Conducted by Students

Workshops Created and Taught by Students

Data Science Student Consultants

Flynn Chen, Medicine
Tuba Dohlar, Mechanical Engineering
Daniel Encinas Zevallos, Political Science
Arne Holverscheid, Political Science
Isabel Knight, Data Science Undergraduate
Yangyang Li, Driskill Graduate Program in Life Sciences
Jenny Liu, Medicine
Julie Anh Nguyen, Applied Math
Yaelle Pierre, Data Science Undergraduate
Anthony Pulvino, Interdisciplinary Biological Sciences
Carrie Stallings, Sociology
Elizabeth Teng, Astronomy
Daniel Turner, Linguistics
Samvardhan Vishnoi, Physics
Eileen Zheng Wu, Psychology
Patrick Zacher, Cognitive Psychology

Staff

Team members and their positions during the 2023-2024 academic year:

Christina Maimone: Associate Director, Research Data Services
Colby Witherup Wood: Manager, Data Science Services
David Nichols: Lead Statistician
Aaron Geller: Senior Data Visualization Specialist
efrén cruz cortés: Data Scientist
Ritika Giri: Data Scientist
John Lee: Data Scientist
Emilio Lehoucq: Data Scientist
Jillian Whitton: Statistician