NCSI: Predicting Eviction Events Based Upon Utility Bills

Status	Completed
Mentor Name	J Bryan Osborne
Mentor's XSEDE Affiliation	Faculty Researcher
Mentor Has Been in XSEDE Community	Less than 1 year
Project Title	Predicting Eviction Events Based Upon Utility Bills
Summary	The City of Tulsa has challenged the ORU Data Science Team to analyze several years of utility bills and eviction actions, and develop predictive models for residential eviction likelihood in order to trigger a potential intervention in order to avoid actual eviction if possible. Existing attempts to address this problem have not resulted in satisfactory predictability.
Job Description	Students will be tasked with data reformatting, cleaning, annotation, etc., to prepare for model creation and training. Faculty-led efforts will allow students to develop multiple machine learning models compare predictability results. Stretch goal will be to field the model operating on real-time data in a test environment provided by the City of Tulsa.
Computational Resources	Project will use the ORU Research Computing and Analytics resources including the Titan Supercomputer. If additional resources are required, allocation will be sought from University of Oklahoma OSCER and/or XSEDE National Center resources.
Contribution to Community	The City of Tulsa continues to struggle with some of the highest eviction rates in the country. The high rate of evictions has existed for several years (https://bit.ly/3nG6LYC) and COVID has only exacerbated the problem (https://bit.ly/3bifNFs). Evictions continue to lead to disruptions in the lives of Tulsa families and are causing increases in homelessness in the community (https://bit.ly/3Cqo7yW).
Position Type	Apprentice
Training Plan	Participating students will have either completed introductory Data Science courses offered at ORU or will utilize online training to acquire the fundamental skills required to perform the study. Continuing and "just-in-time" education will be provided by the research supervisor in both group and individual settings. Guided learning and a collaborative approach will be utilized. Students will work on specific research tasks assigned by the research supervisor as well as participate in group reviews of research problems, how to best solve them, and results of those efforts.
Student Prerequisites/Conditions/Qualifications	Completion of Introduction to Data Science course at ORU or equivalent. Knowledge of R (and potentially Python). Grasp of basic statistical concepts necessary for data analysis.
Duration	Semester
Start Date	01/17/2022
End Date	05/06/2022