Optum Research Project Specification

Project Theme

“The impact of race, social and demographic factors on health, survival, and mortality.”

The motivation and purpose for the 2023 Optum project is to highlight the issues and external factors that impact people’s mental and physical health, which ultimately determine the quality of one’s life. This project will example opportunities to improve disparities by identifying social, racial, demographic factors that impact and predict quality of mental and physical health.

Your task

Each student has been allocated into a project team (of up to 4 students). Each project team has been assigned a specific project research topic using the above general theme as a prompt. Your goal is to complete the required project deliverables for your assigned project, in accordance with the guidelines detailed in the remainder of this document.

Project deliverables

This project has the following three key deliverables, due in the final week (starting 07/241):

  1. Reproducible project report

    • This report is to be written as either an Rmarkdown or a quarto markdown file.

    • Submitted as rendered pdf and html output formats.

    • The report must comply with the rubric provided.

  2. Project poster

    • Required to be an A2 landscape size.

    • Submitted as a pdf output format, and a printed copy for the poster session.

  3. Project slide deck and presentation

    • Executive pdf output summary of key poster results (maximum 5 slides).

    • Used in a 5 minute presentation by the entire team on the final day.

It is important to know that the project poster is summarized directly from the project report, and the slide deck further summarizes the key insights from both the project report and poster. The project report thus contains all of the detailed reproducible analysis for your research question. However you are expected (and advised) to work on both the project report and poster concurrently to meet the weekly project milestones.

All three project deliverables must only use open data (see data guidelines), and will be publicly linked on our showcase website. The project report analysis must be made available online. It must be fully reproducible and developed only using open source software as taught in the course.

Benefits of the Program

These three deliverables are designed to be openly accessible research projects. The main benefits to you in completing them successfully include (but are not limited to) the following:

  1. Networking opportunities: you will actively present your poster and slides to Optum Executives and CMU Statistics & Data Science faculty.

  2. Building a public data science portfolio: your report, poster, and slide deck will be publicly shared as part of our program showcase.

  3. Developing marketable skills: the project involves collaborative software skills essential to industry and academia (version control, reproducibility), in addition to the statistical techniques you learn. Most importantly, you will practice communicating your work to a broad audience.

Producing high quality project deliverables is an ideal way to broadcast your research for future health and data science related job-search purposes. So please put your best foot forward!

Research Topic Guidelines

Using the general project theme and the required data guidelines, your project team has been assigned to one of the following 4 research topics.

Final Research Topics

  1. Premature deaths.

    • TA Advisor: Akshay Prasadan

    • Do income inequality, unemployment and high school completion rates affect the number of premature deaths of certain racial groups at the county level?

  2. Preventable hospital stays:

    • TA Advisor: Beomjo Park

    • Do income inequality, unemployment and high school completion rates affect the number of preventable hospital stays of certain racial groups at the county level?

  3. Mental health:

    • TA Advisor: YJ Choe

    • Do the number of mental health professionals per county affect the number of poor mental health days?

  4. Drug overdose alcohol related deaths:

    • TA Advisor: Alec McClean

    • Are there demographic and social factors that are predictors of drug overdose, alcohol-related incidents (e.g., driving accidents)?

Exploratory and Predictive modeling

In preparing your three project deliverables, you should focus on both:

  • Exploratory analysis: Create visualizations to explore the underlying structure of key demographics and social determinants in the data, gaining insight about distributions and relationships between variables. These should be ideally based on reasoned hypotheses by the team.

  • Predictive modeling: Are there demographic and social determinants, access to care factors that are predictors of physical and emotional health. These could be the outcome of carefully applied predictive models.

Data Guidelines

Required: County Health Rankings Data

In order to complete this project successfully, you must utilize (and appropriately cite) relevant datasets as detailed below:

  • Required data source: The County Health Rankings Data

  • The County Health Rankings Data as collected by the University of Wisconsin Population Health Institute, ranks every county in each state on their Health Outcomes and Health Factors. This dataset also contains the measurements used to calculate the rankings for each county.

  • We recommend reading the background information and video explainers for details.

  • In particular you must (at minimum) utilize the 2023 ranking measures which can provide more insight into the most recent health ranking outcomes priorities. This can be used to better shape your project topic and related hypotheses.

  • Your analysis should mainly be done at the entire United States scale (as feasible). However you are welcome to focus on some specific counties/states to test more granular spatial hypotheses.

  • All data used must be publicly available. Any personally identifiable data, or data that is not openly accessible must be removed from your project deliverables before submission.

Optional: Additional Suggestions and Data Sources

The following are some additional guidelines and data sources which you can consider during the course of your project analysis:

  • The County Health Rankings Data are typically collected over time.

    • Consider doing a temporal or trend analysis for your analyses.

    • For predictive modeling consider adding time-varying features or forecasting an outcome, with suitable uncertainty quantification.

  • Consider whether this County Health Rankings Data can be augmented with other publicly available datasets, over the time periods.

    • Consider spatiotemporally merging on US Census, or COVID19 data to the County Health Rankings Data.

    • Such data is publicly accessible in R using the {tidycensus} and {covidcast} packages, for example.

    • We note that the County Health Rankings Data is a requirement, but these optional augmented datasets can help you test, and sharpen your hypotheses for some of the collected ranking metrics across US Counties.

Project Logistics

Project Teams and TA advisor

  • Based on your pre-course Optum survey topic preferences, each student has been assigned to a project team of (up to) four students.

  • Your project team has already been assigned to a research topic.

  • You will maintain this assigned team structure and project topic throughout the program.

  • TA Advisor: Your project team will be assigned a TA advisor. Your assigned TA advisor will be your project teams’ primary point of contact throughout the program.

  • All students however are expected to do the majority of the project work together, including collecting datasets, formulating hypotheses, EDA, and predictive modeling etc.

  • The TA advisor will help project guidance and mentorship and mainly help resolve any high level technical issues during your project work.

  • All “project-lab” activities throughout the program will involve meeting with your assigned project team members and typically the TA advisor.

  • It is important for your team members to meet during the assigned times and ensure that you meet the weekly project milestones.

  • Student teams are encouraged to meet during free times during the workday, outside of lectures to continue making progress on the final project.

Project Milestones

These project milestones are to be met the Friday of each week. During your Friday project lab session, the head instructor (Shamindra Shrotriya) will briefly meet with your project team and your TA advisor. During this meeting, you will provide quick progress update on that weeks’ project milestones, as detailed below:

  • Week 1:

    • Meet your team and TA advisor.

    • Start to understand your topic of interest and start noting formulating specific hypotheses.

    • Your TA advisor will help you set up a collaboration workflow (e.g., Dropbox, Github).

    • Set up the project report document as either a Rmarkdown or a quarto markdown file, and start populating it.

  • Week 2:

    • Formulate initial key hypotheses with TA advisor.

    • Source initial datasets and read them into R.

    • Do some basic exploratory data analysis (EDA), i.e., produce some summary plots and update your project report. This may change, but keep adding in findings.

    • TA Advisor to set up an overleaf project poster file for the team. Must be A2 landscape format. Your team can use other formats per your preference.

    • Start also populating some poster sections like the introduction, data sources, and placeholder for the conclusion. Mirror these from your project report.

    • Your project report and poster should be done concurrently to save time!

  • Week 3:

    • More hypotheses should be tested and introduction, data sources section should be finalized in both the project report and project poster.

    • Additional exploratory plot with commentary should be finalized in both the project report and project poster.

    • Carefully list assumptions, not just insights when summarizing any findings.

    • Start preparing predictive modeling questions of interest and source some datasets in R to help answer the questions.

  • Week 4:

    • Additional exploratory plots with commentary should be finalized in both the project report and project poster.

    • Predictive models should be run. Check which predictors provide most explanatory power for the outcome. Note down key assumptions and data limitations as you draft your report and poster findings.

  • Week 5:

    • Refine your predictive models. Draft predictive questions section in the poster, including question, data, and methodology used. Explain your model and variable selection procedures used.

    • Report should be drafted at this stage, similar to the poster.

  • Week 6:

    • Polish the report and poster. Ensure that you can generate pdf and html files for the report submission.

    • Create a basic slide deck and summarize main results. This should be a quarto markdown reveal.js document, as taught in the course lectures.

    • You will have only 5 minutes for your presentation, so it should be a maximum of 5 slides.

  • Week 7:

    • Optum HQ visit 07/17 to 07/20 (inclusive).

    • Not much work to do this week.

    • Just keep polishing the report and poster session in the background on the Friday project-lab.

  • Week 8:

    • Submit your project poster pdf by 07/27 to allow for printing.

    • Submit your project report pdf and html files by 07/27.

    • Submit your project slide deck pdf by 07/27.

Footnotes

  1. Details on deliverable due dates will be announced after semester starts. Stay tuned!↩︎