3 Levels of a DH Project

Selecting Our Sources

Our project combines data from multiple sources to analyze healthcare experiences across U.S. states. The primary dataset, sourced from the Centers for Medicare & Medicaid Services (CMS), includes variables such as star ratings for patient care quality and Medicare spending per episode of care by state. Collected in late 2024 and released in early 2025, this dataset aims to showcase the quality of home healthcare providers. However, its governmental funding introduces potential biases, such as the exclusion of Puerto Rico and most U.S. territories, limiting inclusivity. Additionally, the undisclosed wording of patient satisfaction survey questions raises concerns about framing biases that could influence responses. These silences, as discussed in D’Ignazio and Klein’s Data Feminism, highlight how societal power dynamics can marginalize certain voices in data collection.

To complement the CMS data, we incorporated chronic respiratory disease rates from 2021, sourced from the Institute for Health Metrics and Evaluation (IHME) and the Global Health Data Exchange (GHDx). Funded by independent research organizations focused on public health, this dataset appears relatively unbiased. However, it has limitations: the 2021 data may be skewed by the COVID-19 pandemic, which disproportionately affected individuals with chronic respiratory conditions. Additionally, the dataset covers the general population, not just Medicare beneficiaries (aged 65+), which could lead to discrepancies when comparing states with varying age demographics.

We chose these datasets because they align with our research focus on healthcare delivery and patient outcomes. The CMS dataset provides insights into patient satisfaction and healthcare processes, while the chronic respiratory disease dataset offers measurable health outcomes, enabling us to explore correlations between healthcare delivery and patient health. To contextualize our analysis, we also incorporated historical Medicare policy data from sources like the Kaiser Family Foundation’s Medicare Timeline and peer-reviewed articles on Medicare’s evolution. These sources helped us understand how policy changes have shaped patient experiences over time.

PROCESSING OUR DATA

The healthcare experience dataset was relatively clean, requiring minimal preparation for analysis. In contrast, the chronic respiratory disease dataset needed more extensive processing due to its broad scope across multiple years. We focused on the most relevant column: the probability of death from chronic respiratory diseases in 2021, selected for its recency and alignment with our primary 2024 dataset.

Using the Global Health Data Exchange (GHDx) website, we filtered and extracted state-level chronic respiratory disease death rates. We then merged this data with the healthcare experience dataset using state identifiers in Google Sheets, enabling a state-level analysis of the relationship between patient satisfaction, healthcare delivery, and chronic respiratory disease outcomes.

For data visualization, we primarily used Tableau, leveraging its interactive and versatile capabilities to explore relationships between variables and present complex data clearly. Additionally, we used Timeline.js to create an interactive timeline of key Medicare policy milestones, integrating visuals and descriptions to make historical events engaging and accessible. Together, these tools allowed us to present our findings in an informative and visually compelling manner.

Presenting Our Narrative

We designed our website using WordPress. To align with our healthcare-focused topic, we selected the Variations theme by Tyler Moore and implemented a blue color scheme. To enhance the historical context and visual appeal, we incorporated video footage of Former President Harry S. Truman signing the Social Security Amendments of 1965, which established Medicare. Featured prominently on the homepage, this footage underscores the significance of Medicare’s creation and sets a compelling tone for our narrative. The clean design, thematic color scheme, and impactful visuals work together to engage users and effectively communicate the importance of our research.