Summer 2015

Robin Dunn*, Department of Mathematics, Kenyon College
Supervisor: Eddy

Kelli Marquardt, Department of Mathematics, University of Dayton
Supervisor: Eddy

Brandon Sherman*, Department of Statistics, University of Pittsburgh
Supervisor: Eddy

Summer 2014

Kevin Eng, Department of Statistics
Supervisor: Eddy

Nick Ettlinger, Department of Statistics
Supervisor: Eddy

Summer 2013

Kairavi Chahal, Department of Statistics
Supervisor: Steorts

Nicole Crimi, Department of Statistics
Supervisor: Eddy

Emily Furnish*, Science and Humanities Scholars (International Relations & Politics)
Supervisor: Fienberg, Steorts

Dahiana Jimenez, Department of Statistics
Supervisor: Steorts

Summer 2012

Brittany Buggs*, Department of Statistics, Heinz College Public Policy
Supervisor: Eddy

Nicole Crimi, Department of Statistics
Top Coding in the US Census and American Community Survey[.pdf]
Supervisor: Eddy

Sepideh Mosaferi*, Dietrich Humanities & Social Sciences
Supervisor: Fienberg

Jaime Trujillo, Department of Statistics
Errors in Census and American Community Survey Data File[.pdf]
Supervisor: Eddy

George Volichenko*, Department of Statistics
Capture-Recapture Interactive Simulation Package[.pdf]
Supervisor: Eddy

Abbas Zaidi*, Department of Statistics
Supervisor: Eddy

Olga Zubashko, Department of Statistics
Supervisor: Eddy

* denotes a student who continued to graduate or professional school

These research projects are typically one or two semesters long.
List does not include Senior Honor Thesis work (see menu at left).

AY 2015-2016

Ernest Chiew*, Science and Humanities Scholar, Departments of Physics, Statistics
Supervisors: Eddy, Ventura

AY 2014-2015

Nick Ettlinger, Department of Statistics
Supervisor: Eddy

AY 2013-2014

Nick Ettlinger, Department of Statistics
Supervisor: Eddy

Shannon Gallagher*, Department of Mathematical Sciences
Supervisors: Fienberg, Eddy

AY 2012-2013

Nicole Crimi, Department of Statistics
Top Coding in the US Census and American Community Survey [.pdf]
Supervisor: Eddy

Michael Pane*, Department of Statistics
Struggles in Small Area Estimation: Benchmarking and Weighting
Supervisor: Steorts

Olga Zubashko, Department of Statistics
Supervisor: Eddy

* denotes an undergrad who continued to graduate school

Dietrich College of Humanities & Social Science Honors Thesis Program (Link)

Academically qualified students are invited by the Dietrich's Dean Office to participate in the Honors Thesis Program. Students then align their research interests with a deparment and supervisor. Although a large group of students may be invited on the basis of their GPA, in practice the Statistics Department chooses only 3-5 top students a year. Below are students whose theses are grounded in the research problems of this node.

2014-2015

Popov, M (2013). Visualizing the 2010 US Census Data with Shiny. [.pdf] (2013).
Full source code and instructions at GitHub.

Thomas, A.C., Ventura, S., Popov, M., and Nugent , R. (2013). Visualizing the Census and American Community Survey: An Education Module for Undergraduates. [.pdf].

This course is an introduction to "data matching", the field of improving data quality through record linkage techniques including merging lists, identifying duplicates, and unique entity resolution. We discuss working with text and its similarity metrics, imputation models for missing data, statistical models for predicting matches, and real-life case studies of record linkage applications. The course is designed for a mini (about 7 weeks) for advanced undergraduates and master's level students.

Instructors: Steve Fienberg, Rebecca Nugent, Department of Statistics (both Node members)
TA: Sam Ventura, Department of Statistics (Node member)
13 students

Textbooks: Data Quality and Record Linkage Techniques by Herzog, Scheuren, Winkler (Springer, 2007); Data Matching: Concepts and Techniques for Record Linkage, Entity Resolution, and Duplicate Detection by Peter Christen (Springer, 2012)

Syllabus

Lectures:

Lecture 1: Introduction; Overview
Lecture 2: Motivation; Case Studies; Data Structure
Lecture 3: Text Similarity Metrics; Supervised Models; A Comparison of String Distance Metrics for Name-Matching Tasks (Cohen, et al); Adaptive Name Matching in Information Integration (Bilenko, et al)
Lecture 4: Deterministic vs Probabilistic; Evaluation; ROC Curves; Microsoft Case Study
Lecture 5: Supervised Model Review; Logistic Regression, Classification Trees; infection.txt
Lecture 6: Fellegi-Sunter Model; A Theory for Record Linkage (Fellegi, Sunter, 1969)
Lecture 7: Fellegi-Sunter Model; E-M Algorithm; Transitivity; Single Linkage Hierarchical Clustering
Lecture 8: E-M Algorithm; Logistic Regression for Fellegi-Sunter; Determining Record Linkage Parameters using an Iterative Logistic Regression Approach (Harville and Moore)
Lecture 9: Blocking; Capture-Recapture; Handout from Lohr
Lecture 10: Transitivity; Clustering; Linking more than Two Lists; A Generalized Fellegi-Sunter Framework for Multiple Record Linkage with Application to Homicide Record Systems (Sadinle, Fienberg, 2013)
Lecture 11: Record Linkage when there's too much data

Homeworks:

HW 1 Winkler 1990 string comparison paper
HW 2
HW 3 list1.txt list2.txt list3.txt list4.txt

Project Information and Examples:

Handouts: Brainstorming Project Ideas; Presentation Handout; Report Handout
Examples:
String Similarity Metrics Comparison for Name-matching Task
Matching Migration Records: Record Linkage and Capture-Recapture Analysis of Kosovo Border-Crossing Records in 1999
Similarity Metric and Nonmetric Scores: A Comparison
Image Recognition Using Edit Distance
Missing Data Problems in Record Linkage - How to Find Links and Non-Links
Matching Artists Names from Two Sources: A Record Linkage Application
The Soundex Algorithm's Performance in Record Linkage and De-duplication
Clustering Linkage Functions in Record Linkage

This course is an introduction to privacy and confidentiality and is designed as a short course for professional master's students. The course was given at the University of Paris Dauphine. All materials are authored by Rebecca Steorts.

Course details can be found here .

Statistical Graphics & Visualization is a 300-level elective class for sophomores, juniors, and seniors in Statistics, Information Systems, Computer Science, Math, and other majors interested in visualizing data. This course also satisfies the Creating General Education requirement for Dietrich College and is highly oversubscribed with a long waitlist every year.

Students learn how to extract, summarize, and visualize structure and features for low- and high-dimensional data. The class is taught in R, but other languages could be subsituted. For the final project, teams of students develop hypotheses and create statistics graphics to look for supporting or contradictory evidence. Projects are presented in poster or website form.

Spring 2017

Instructor: Sam Ventura, Department of Statistics (Node member)
TAs: none were node members this year
99 students

This version included modules on working with geo-spatial and census data.

Fall 2016

Instructor: Sam Ventura, Department of Statistics (Node member)
TAs: included Kayla Frisoli (Node member), Department of Statistics
65 students

This version included modules on working with geo-spatial and census data.

Spring 2016

Instructor: Sam Ventura, Department of Statistics (Node member)
TAs: included Kayla Frisoli (Node member), Department of Statistics
106 students

This version included modules on working with geo-spatial and census data.

Spring 2013

Instructor: Andrew Thomas, Department of Statistics
TAs: Sam Ventura and Mikhail Popov, Department of Statistics (both Node members)
55 students

Visualize the distributions of population characteristics in the metropolitan area(s) of interest using data sets from the US Census; Project Instructions/Handout: pdf

(Presented at NISS, 2013)
Summary of Module: poster
Working with Shiny and R: poster

Shiny applications are hosted on the Department's server using RStudio's Shiny Server software.

Undergraduate Research is a small invitation-only statistical research class for top juniors and seniors in statistics, machine learning, mathematics, computer science, etc. Interested students apply in the fall; between 12-15 students are selected. Small groups of three or four students are paired with a faculty client often in another discipline whose project requires statistical expertise and methodology. Students work all semester on this project with feedback and guidance from both the course professir and the faculty client. The goal is to expose students to the research process. Students present their work in both report and poster format with the latter being at CMU's Meeting of the Minds Undergraduate Research Symposium. The below are projects aligned with the research of this node.

Spring 2014

Instructor: Cosma Shalizi, Department of Statistics
TA: Beau Dabbs, Department of Statistics

Deduplication of Civil War Killings in El Salvador
Daniel In, Karn Mishra, Joseph Pane
Faculty client: Steorts

Sampling, Survey, and Society is an elective for majors in Statistics, Economics, and other students in the Dietrich College of Humanities and Social Sciences, drawing students with a very wide variety of backgrounds. This course revolves around the role of sampling and sample surveys in the context of U.S. society and its institutions. We examine the evolution of survey taking in the United States in the context of its economic, social, and political uses, eventually leading to discussions about the accuracy and relevance of survey responses, especially in light of various kinds of sampling errors. Students are required to design, implement, and analyze a survey sample and then write a scientific report summarizing their survey and its results.

Spring 2017

Instructor: Jared Murray (Node member), Department of Statistics
TAs: included Nicolas Kim (Node member), Department of Statistics
71 students

Spring 2016

Instructor: Jared Murray (Node member), Department of Statistics
TAs: included Nicolas Kim (Node member), Department of Statistics
61 students

Spring 2015

Instructor: Jared Murray (Node member), Department of Statistics
TAs: Maria Cuellar, Nicolas Kim (both Node members), Department of Statistics
44 students