Tag: Data Science

Detecting Business Communication Patterns Using Machine Learning and LLMs (Lingshu Hu)

Remote: this project is scheduled to begin on 6/2/2025 and run for 10 weeks, finishing on 8/8/2025.

Project Description

This project consists of two studies. The first study explores how DEI-related content shared by firms on Twitter (now X) influences the digital engagement of their customers. Previous research has shown that brand-generated content on social media plays a critical role in marketing success, and our prior study indicates that DEI-related content generally enhances brands’ digital engagement, with this effect moderated by factors such as the political spectrum of their followers and the type of firm. This new project seeks to further investigate the impact of firms’ DEI strategies on Twitter, focusing on variables such as the authenticity, relevance, and emotional-cognitive aspects of DEI-related content. It also examines whether these strategies affect firms’ financial performance. To support this research, we have developed machine learning models to detect DEI-related content and assess the political spectrum of firm followers. We will also develop new models to measure more nuanced concepts such as authenticity and relevance, as well as econometric models to analyze the effects of these communication patterns. The second study investigates how large language models (LLMs) can predict the outcomes of oral communication between sales representatives and customers. It aims to uncover effective strategies in business communication by applying LLMs to identify key cues for successful conversations. Additionally, we plan to build a system to enhance the communication strategies of sales representatives.

Apply now

Prerequisites

BUS 202 or equivalent

Special Comments

Project Information (subject to change)

Estimated Start Date: 6/2/2025

Estimated End Date: 8/8/2025

Estimated Project Duration: 10 weeks

Maximum Number of Students Sought: 1

Research Location: Remote

Contact Information: Lingshu Hu (email: lhu@wlu.edu)

CLOSED: Green Information Systems Comparative Study (Keri Larson)

Remote: this project is scheduled to begin on 6/2/2025 and run for 10 weeks, finishing on 8/8/2025.

Project Description

This project focuses on international Green IS, or sustainable information systems, efforts. Specifically, the student research scholar will assist investigation into the sustainability practices of two nations, Iceland and Wales, that are considered leaders in the global effort to create a viable low-carbon world. The project will comprise both qualitative and quantitative data collection including assessments of both governments’ legislative endeavors and adherence to the UN’s Sustainable Development Goals. Data will be collected in a low-tech environment, so the major requirement is familiarity with Excel. Depending on the trajectory of the project, some data will be collected via survey. As such, the student research scholar will gain experience in survey design methodology if desired, entailing the use of the Qualtrics survey platform.

Apply now

Prerequisites

Familiarity with Excel Willingness or interest in learning survey design and learning to use Qualtrics

Special Comments

Project Information (subject to change)

Estimated Start Date: 6/2/2025

Estimated End Date: 8/8/2025

Estimated Project Duration: 10 weeks

Maximum Number of Students Sought: 2

Research Location: Remote

Contact Information: Keri Larson (email: larsonk@wlu.edu)

Unraveling the Spider’s Circadian Clock (Natalia Toporikova)

On campus: this project is scheduled to begin on 6/2/2025 and run for 8 weeks, finishing on 7/25/2025.

Project Description

Are you fascinated by the intricate workings of nature’s timekeepers? Do you have a passion for interdisciplinary research? Are you ready to spin a web of knowledge about arachnid chronobiology? If so, we have an exciting summer research opportunity for you! We anticipate 2 students joining our dynamic lab team to explore the biological mechanisms of the spider’s circadian clock. This cutting-edge research project aims to show how these eight-legged marvels regulate their internal rhythms. We welcome applications from students across various majors(or inspiring majors), including: Computer Science( Develop websites to conduct data analysis), Physics (Create models of circadian systems), Engineering (Signal processing of spider locomotion activity), Neuroscience (Design and conduct spider behavior experiments), Data Science (Analyze complex behavioral datasets) ï Biology: Investigate gene expression patterns related to circadian rhythms

Apply now

Prerequisites

Appreciation for spiders and eagerness to learn more about them. Ability to work collaboratively in a diverse team. Commitment to learn Python programming language

Special Comments

To accomplish the project, you must learn how to work with data frames in Python (pandas library). If you are not familiar with pandas, you need to take either Biol-187, Intro to Data Science (Winter 2025) or Biol-297, Behavioral Data Science (Spring 2025)

Project Information (subject to change)

Estimated Start Date: 6/2/2025

Estimated End Date: 7/25/2025

Estimated Project Duration: 8 weeks

Maximum Number of Students Sought: 2

Research Location: On campus

Contact Information: Natalia Toporikova (email: toporikovan@wlu.edu)

CLOSED: Comparing AI Music to Human Composed Music using Benford’s Law (Sybil Prince Nelson)

On campus: this project is scheduled to begin on 6/9/2025 and run for 6 weeks, finishing on 7/18/2025.

Project Description

Benford’s Law is a mathematical principle that predicts the distribution of leading digits in naturally occurring datasets. It’s commonly used for detecting anomalies in financial records and identifying fraudulent behavior, such as social media troll accounts. This summer, our research lab will explore novel applications of Benford’s Law in the analysis of music through three distinct projects: Project 1 will focus on comparing AI-generated music to human-composed music, using Benford’s Law to determine whether statistical differences exist in their digit distributions. This could potentially reveal unique patterns inherent to human creativity versus algorithmic composition. Project 2 will create a comprehensive database of songs, including metadata like genre, artist, and Benford-related statistics, to serve as a foundation for future research in music analytics. Project 3 will develop a simulation that generates original music that conforms to Benford’s Law. This involves building algorithms capable of composing music that statistically aligns with the expected leading digit patterns. Students involved will gain experience in statistical modeling, programming, and music data analysis, bridging the gap between mathematics, computer science, and music. No prior experience is requiredójust curiosity and a desire to explore the creative possibilities at the intersection of these disciplines.

Apply now

Prerequisites

Knowledge of music and/ or computer programming will be helpful.

Special Comments

There is the potential to attend a conference in November 2025 in South Korea depending on the results of this research.

Project Information (subject to change)

Estimated Start Date: 6/9/2025

Estimated End Date: 7/18/2025

Estimated Project Duration: 6 weeks

Maximum Number of Students Sought: 4

Research Location: On campus

Contact Information: Sybil Prince Nelson (email: sprincenelson@wlu.edu)

Using 2nd and 3rd generation generative AI to discover novel approaches to genome annotation (Gregg Whitworth)

On campus: this project is scheduled to begin on 6/2/2025 and run for 10 weeks, finishing on 8/8/2025.

Project Description

All modern biomedical studies conducted at a systems level depend on high quality genome annotation. But there are a very small number of organisms for which high quality genome annotations exist. This problem frustrates work in both model research organisms and in mammalian genomes closely related to humans. Efforts were made 20 years ago to use ML techniques to pseudo automate the annotation of genes, most of which largely failed. Recent advances in LLM-based approaches suggest the time may have come for this approach to finally be useful and productive. I am interested in recruiting new students to work on both the computational project and bench validation.

Apply now

Prerequisites

No. Programming skills in any language are preferred if you are interested in a computational focus.

Special Comments

I recommend all new research students enroll in Biol-401 this winter. We will meet once a week for a few hours to discuss recent papers and possible project directions.

Project Information (subject to change)

Estimated Start Date: 6/2/2025

Estimated End Date: 8/8/2025

Estimated Project Duration: 10 weeks

Maximum Number of Students Sought: 2

Research Location: On campus

Contact Information: Gregg Whitworth (email: whitworthg@wlu.edu)

CLOSED: Machine Learning Models of Beliefs, Values, and Expectations (Jon Eastwood)

On campus: this project is scheduled to begin on 6/16/2025 and run for 6 weeks, finishing on 7/25/2025.

Project Description

This summer’s research will build on research done with SRS students last summer, laying the groundwork for my new project on cynicism and distrust in public life.  This summer, we will use a variety of machine learning models to explore patterns of distrust and cynicism in global public opinion, focusing on World Values Survey data (https://www.worldvaluessurvey.org), joined with country and country-year data from a variety of sources (UN Human Development Reports, the World Bank, and more).  This will include using random forest models to predict individuals’ trust responses and then using these models to select important features for explanatory modeling.  We will also use generalized additive models (GAMs) to explore patterns of interest to the PI and to student researchers, as appropriate.  Questions to explore include the relationship between measures of institutional functioning and (dis)trust, population heterogeneity and (dis)trust, and economic performance and (dis)trust).

Apply now

Prerequisites

Some prior experience working with applied data science (R or Python) is expected.  It would be ideal if student researchers already have some fluency with basic wrangling and data visualization using the tidyverse suite in R.

Exposure to social science theory (e.g., sociology, anthropology, political science, economics) would be helpful.

If applicants have questions about their preparation in these respects they are encouraged to reach out to the professor.

Special Comments

Students with otherwise strong credentials but who don’t have prior R experience could satisfy this requirement by enrolling in the professor’s 1-credit “R for Social Scientists” course in winter, 2025.

Project Information (subject to change)

Estimated Start Date: 6/16/2025

Estimated End Date: 7/25/2025

Estimated Project Duration: 6 weeks

Maximum Number of Students Sought: 3

Research Location: On campus

Contact Information: Jon Eastwood (email: eastwoodj@wlu.edu)