Undergraduate research in statistics provides opportunities for gaining experience in data analysis, reading and writing about statistics, and collaboration with Statistics faculty mentors and their research teams. By doing an undergraduate research project, you will develop a deeper understanding of statistics, whether as a first/second year student considering a statistics major, or as a junior/senior considering graduate school and other career options.
The two largest programs for undergraduate research in statistics are the honors thesis, for juniors & seniors, and UROP, for first & second year students. In addition, some faculty research projects involve undergraduates not in either of these programs. Other statistics research-related activities involving undergraduates include the following:
- The annual Michigan Student Symposium for Interdisciplinary Statistical Sciences (MSSISS). MSSISS provides a forum for presenting completed research projects, and an opportunity to see the range and scope of statistical activity across the University of Michigan. Most of the research projects are carried out by graduate students, but undergraduates are welcome to participate and many have!
- The Statistics department occasionally runs a data mining competition.
- A relevant national forum is the free Electronic Undergraduate Statistics Research Conference, and the associated Undergraduate Statistics Project Competition.
- The Center for Statistics, Computing, and Analytics Research (CSCAR) sometimes employs undergraduates. Email email@example.com if you are interested in learning more about opportunities for involvement with CSCAR.
Writing an Honors Thesis
An honors thesis provides an opportunity for eligible students to carry out faculty-supervised research in the senior year. The application process and requirements for the Statistics, Data Science, and Informatics honors programs are described on the department website. Students are encouraged to contribute their thesis to the archive of honors theses at the University of Michigan Library.
Past Honors Theses
- Chenxi Fan, Statistics - An evaluation of information criteria for model selection in quasi-likelihood regression, with application to modeling COVID mortality and case incidence in the United States
- Siqi Li, Statistics - Local False Discovery Rates in the Multi-Parameter Case, with Application to Epigenetics of Human Growth
- Wanqi Liang, Data Science - An Applet and Tutorial for Calculating the Sample Size (and Power) for a Clustered Sequential, Multiple Assignment Randomized Trial
- Juejue Wang, Statistics - Comparison of Document Co-clustering aslgorithms and Application of Single-cell RNA-seq Data Clustering to Twitter Data
- Chao Peter Yang, Data Science - The Classical-Romantic Dichotomy: A Machine Learning Approach
- Ziyang Shao, Statistics - College Ranking Based on Pairwise Preferences
- Haoyu Chen, Statistics - Kernel Methods for Activation Energy Prediction
- We Han, Statistics - Argo Data Mean Field Modeling
- Jiahui Ji, Statistics - NYC Optimal Transport and Ridesharing
- Xiaotong Yang, Statistics - Fitting mechanistic models to Daphnia panel data within a panelPOMP framework
- Shuaiji Li, Statistics - Auto Sales Prediction with attention to the Parable of the boiled frog: Functional Data Analysis and Time Series Forecasting
- Zifan Li, Statistics - Perturbation Algorithms for Adversarial Online Learning
- Tianwen Ma, Statistics - A Functional Data Analysis Approach to Looking at Handwriting Data
- Xige Zhang, Statistics - Robustness of the Contextual Bandit Algorithm to A Physical Activity Motivation Effect
- Rong Zhou, Statistics - The Comparison of ACI and MCB Methods for Choosing a Set that Contains the Optimal Dynamic Treatment Regime
- Xinyan Han, Statistics - An Empirical Comparison of Various Online Binary Classification Algorithms
- Hwanwoo Kim, Statistics - A Sample Size Calculator for SMART Pilot Studies
- Yuchen Lin, Statistics - Auto Car Sales Prediction: A Statistical Study Using Functional Data Analysis and Time Series
- Kelsey Pakkala, Statistics - A Functional Data Analysis Approach to Women’s Health Screening Adherence for Breast Cancer and Cervical Cancer
- Emily Slade, Statistics - Functional Data Analysis in Cephalometric Tracing and Mandibular Examination
- Ben Charoenwong, Statistics - An Exploration of Simple Optimized Technical Trading Strategies
- Matthew Lomont, Statistics - Detecting Active Pathways in Gene Sets
- Xuanzhong Wang, Statistics - An Exploration of Influential Observations in the Panel Study of Income Dynamics - An Exploration of Gender Gap in Labor Market; Money Resource Allocation to Children in PSID
- Christopher Worsham, Statistics - A Stochastic Model of Retinal Development in Zebrafish
Undergraduate Research Opportunity Program (UROP)
UROP is a great way to get an introduction to research during the first two years at University of Michigan. See the UROP website for more information. For the most part, Statistics and Data Science research projects require foundational preparation in statistics, mathematics and computer programming. Sometimes, first year students have sufficient preparation through AP courses and other experiences. Otherwise, it may be appropriate to take introductory statistics, computer programming and calculus courses in the first year to be ready for a second year UROP project.
Other Opportunities for Undergraduate Research
It is possible to conduct undergraduate research that does not fall into either the honors program or UROP. If you find yourself interested in the research agenda of a Statistics faculty member, you can email to enquire about available options. This research can be carried out as part of Stats 489 [Independent Study in Statistics], as a paid position if one is available, or as an informal arrangement for neither course credit nor payment. Arrangements must be made on a case-by-case basis with the potential faculty superviser.
Faculty Supervising Undergraduate Research
• Danny Almirall supervises undergraduate researchers with an interest in applied issues in causal inference, dynamic treatment regimens and sequential multiple assignment randomized trials (SMART). Projects include:
o Topics in design and analysis of clinical trials for adaptive treatment plans, by Hwanwoo Kim. Co-advised with Ed Ionides. 2nd prize winner in the national Undergraduate Statistics Project Competition.
o Adaptive intervention designs in substance use prevention.
o An Investigation of Predictor for Tailoring Ecological momentary Assessment and Contextual Recall.
o Introduction to Sequential Multiple Assignment Randomized Trials (SMARTs) with Zero Inflated Count Outcomes for the Development of Dynamic Treatment Regimens (DTRs): with application to substance use research.
If you are interested in working with Dr. Almirall, please visit his web page first to see if he is currently accepting new students: http://www-personal.umich.edu/~dalmiral/.
• Moulinath Banerjee has supervised undergraduate projects including:
o Detecting Active Pathways in Gene Sets.
• Ben Hansen has supervised undergraduate projects including:
o Proposals for Generating and Utilizing Well Informed Initialization Values to Improve the Computational Efficiency of Optmatch.
• Xuming He supervises UROP students and advanced undergraduate research in a broad area of statistics. Examples include:
o Monte Carlo evaluation of Value-at-Risk.
o Ordering of multivariate Data.
• Al Hero has supervised undergraduate projects including:
o Dynamic distributed multidimensional scaling (MDS) for data visualization.
o Spatio-temporal network anomaly detection in Abilene data streams.
o Canonical correlation analysis for sunspot and coronal mass ejection image representation.
• Tailen Hsing has supervised undergraduate projects including:
o Analyzing Argo Data Co-advised with Stilian Stoev
o Argo Data Mean Field Modeling Co-advised with Stilian Stoev
• Ed Ionides has supervised undergraduate projects including:
o Topics in design and analysis of clinical trials for adaptive treatment plans. Co-advised with Danny Almirall. 2nd prize winner in the national Undergraduate Statistics Project Competition.
o Modeling cholera as a stochastic process.
o Building POMP objects in R for a dynamic general stochastic equilibrium model..
o Investigating sequential Monte Carlo methods for time series analysis.
o Identification of insurance companies at risk of insolvency. Co-advised with Kristen Moore.
• Long Nguyen has supervised undergraduate projects including:
o Traffic Flow and Density Analysis of NYC TLC Taxi Data.
o NYC Optimal Transport and Ridesharing.
• Kerby Shedden supervises undergraduate research with an emphasis on bioinformatics. Examples include:
o Statistical analysis of high frequency motion capture and muscle activity data: applications to assessing development of trunk postural control.
o Sparsity in the distribution of correlation coefficients in molecular screening data. Co-advised with Ji Zhu.
o Individual-specific and disease-specific factors in acquired copy number variations in cancer.
o Detection of DNA lesions in acute myelogenous leukemia.
o Two-tiered false discovery rates.
o Selective targeting of stem-cell-like cancer cell lines. Co-advised with Gus Rosania.
• Ambuj Tewari has supervised undergraduate research projects and an honors theses. Former projects include:
o Development of an Android app for mobile health.
o Simulations comparing bandit algorithms.
o Development of HeartSteps, an Android app for encouraging physical activity. Co-advised with Predrag Klasnja
o Empirical evaluation of online learning algorithms (honors thesis).
o Numerical experiments with Lasso in high dimensional VAR models.
• Ji Zhu has supervised undergraduate research projects and honors theses. Projects include:
o Forecasting Stock Returns in the Chinese Market with Convolutional Neural Networks.
o Medical Image Classification Building Upon Pre-trained Neural Networks: An Application on Diabetic Retinopathy Detection.