Overview
We'd like a data scientist volunteer to analyze the 20,497 public comments submitted on docket VA-2026-VBA-0067 — a VA proposed rule on evaluative disability ratings and the impact of medication.
Docket on Spicy Regs: https://app.spicy-regs.dev/sr/VA/VA-2026-VBA-0067
Data
A JSON export of the docket (including all comments) is attached to this issue. Each comment includes:
comment (full text)posted_date, receive_date, modify_datecomment_id, title
The comment period ran Feb 17 – Feb 28, 2026 (~11 days), during which 20,497 comments were submitted.
Primary Goals
- Sentiment toward the proposed rule — What proportion of commenters support vs. oppose the rule change? How strong is the sentiment?
- Common concerns/themes — What are the recurring topics and concerns raised by commenters? (e.g., medication access, disability rating fairness, impact on veterans' care, cost concerns)
Open Exploration
Beyond the two primary goals, feel free to explore anything interesting in the data. Some possible directions (not required):
- Are there identifiable commenter groups (veterans, family members, advocacy orgs)?
- Form letters vs. unique personal comments — what's the ratio?
- Did sentiment or themes shift over the comment period?
- Any notable outliers or particularly substantive comments?
Deliverable
A Jupyter notebook or written report documenting:
- Your methodology
- Key findings and visualizations
- Any recommendations or insights
Use whatever tools, languages, and NLP approaches you think are best.
Getting Started
- Download the attached JSON data file
- Explore the data structure (top-level keys:
docket, documents, comments)
- Start with the
comments array — each entry has a comment field with the full text
🤖 Generated with Claude Code
Overview
We'd like a data scientist volunteer to analyze the 20,497 public comments submitted on docket VA-2026-VBA-0067 — a VA proposed rule on evaluative disability ratings and the impact of medication.
Docket on Spicy Regs: https://app.spicy-regs.dev/sr/VA/VA-2026-VBA-0067
Data
A JSON export of the docket (including all comments) is attached to this issue. Each comment includes:
comment(full text)posted_date,receive_date,modify_datecomment_id,titleThe comment period ran Feb 17 – Feb 28, 2026 (~11 days), during which 20,497 comments were submitted.
Primary Goals
Open Exploration
Beyond the two primary goals, feel free to explore anything interesting in the data. Some possible directions (not required):
Deliverable
A Jupyter notebook or written report documenting:
Use whatever tools, languages, and NLP approaches you think are best.
Getting Started
docket,documents,comments)commentsarray — each entry has acommentfield with the full text🤖 Generated with Claude Code