Exploring Organ, Eye & Tissue Donations
Over 100,000 people are on the national waiting list for an organ, eye, or tissue donation.
Despite the need, only 60% of U.S. adults register to be donors after they die.
Could looking at past donor data help us identify which demographics need more outreach in order to increase donor registrations?
Data Source
& Context
The Organ Procurement & Transplant Network (OPTN) publishes data about both living and deceased donors. While the website provides some visualizations, it lacks features that would allow individuals or community organizations to compare how one state or demographic group might compare to another.
Goals
Create an interactive dashboard that enables individuals and organizations to explore donation rates by state, gender, age group or ethnicity.
Analyze rates of donations between living and deceased donors to see if there is a relationship.
Tools
Excel
Python, including Pandas, Numpy, Matplotlib, Seaborn, and Scikit-Learn libraries.
Jupyter Notebooks
Tableau Public
Preparation
Downloaded donor data from OPTN and population data from the U.S. Census Bureau.
Wrangled data to make table structures compatible for merging.
Derived new variables from existing data.
Documented data pre-processing, merging, and population flows.
Analysis
Explored basic statistics and possible correlations between variables using Matplotlib and Seaborn.
Used Scikit-Learn’s machine learning features for linear regression and K-means clustering to investigate a possible relationship between donation rates for living and deceased donors.
Presentation
Created a Tableau dashboard that would allow others to explore this data via interactive choropleth maps, double-bar charts, pie charts, tables and line-charts.
Provided a summary of the statistical analysis comparing donation rates for living and deceased donors.
Results
Visualization of donor data revealed which states and demographic groups have significantly lower donation rates and may need greater outreach.
Visit the interactive dashboard to explore data filtered by year, state, gender, ethnicity, age group or any combination thereof.
Statistical analysis of the donation rates for living and deceased donors confirmed that the two rates operate independently of each other. Neither linear regression nor K-means clustering yielded any statistically significant results.
High Points
I found it incredibly satisfying to wrangle the OPTN and U.S. Census data sets in Python in such a way that they fit together seamlessly.
It was also a treat to work with meaningful real-world data that could potentially impact lives.
Challenges & Lessons
When choosing a data set for this CareerFoundry project, one of the requirements was that it have two quantitative variables. Very little direction was provided as to what might be suitable quantitative data for the types of statistical analysis (and machine learning) that I’d eventually be asked to do.
If I were to approach this project again, I’d ask to see the final rubric in addition to the task directions so that I could chose data that would be more suitable for the types of analysis I’d be asked to do. In other words, make sure the expectations for the final deliverable are clearly defined at the outset of the project!
Final Thoughts
Exploring the data about organ, eye, and tissue donations provided insight into past, but in order to help more people in the future, I’d like to look at the data about the registration rates – where and how are we getting people to sign up to be on donor registries?
Links for further exploration
Organ Donation Vector Drawing adapted from Vecteezy:
https://www.vecteezy.com/vector-art/97098-blood-and-organs-donated