NIH-Budget-Reorg

NIH Budget Alluvial

This repository explores how the recently proposed restructuring of the National Institutes of Health (NIH) fits into decades of historical funding patterns. In March 2025, HHS unveiled the “Make America Healthy Again” reorganization, which collapses the 27 existing NIH institutes and centers into eight new institutes and realigns several functions. We asked: How do those new groupings compare to NIH’s budget trajectory back to FY 2000?

View Results Here

How the Analysis Works

  1. Data Ingestion
    • budgets.csv: Contains historical budgets for each old IC in FY 2000, 2010, 2020, 2024, plus the FY 2026 proposed totals at the new institute level (from the HHS FY 2026 Budget in Brief).
    • mapping.csv: Maps each old IC into its proposed FY 2026 new institute.
  2. Proportional Allocation of FY 2026
    • We sum each new institute’s total FY 2026 request.
    • For each old IC in a group, we calculate
      \text{IC}_{2026} = \frac{\text{IC}_{2024}}{\sum_{\text{group}} \text{IC}_{2024}} \;\times\; \text{NewInstitute}_{2026}.
      
    • This gives every old IC a “destination” value in 2026, so ribbons can flow into the correct new‐institute node.
  3. Building the Alluvial (“Lodes”) Data
    • We combine the four historical years (2000–2024) at the old IC level with the apportioned FY 2026 flows.
    • We order strata so that, in each column, institutes are stacked by their FY 2026 group (smallest → largest), and within each group the old ICs are ordered by their FY 2024 size.
  4. Plotting
    • A single ggplot2 + ggalluvial call renders five vertical slices (2000, 2010, 2020, 2024, 2026), with colored ribbons tracing each IC’s budget over time into its new institute.
  5. Interactive Table
    • We include a DT table of the final lodes data so you can sort, search, and inspect every flow.

Files in This Repo


This work lets you see not only how much each IC has grown or shrunk over the last 25 years, but also exactly how its funding “flows” into the new institute structure proposed for FY 2026.

Attribution

This project was written and coded using ChatGPT o4-mini-high, and fixed by me to actually get it to work. You can read the entire chat here.