After spending the day in Dependency Hell™, today marks the formal release of #VizierDB 2.1.1. The 2.1 branch includes a number of papercut fixes, as well as support for plugins and modularizing workflows as scripts.
https://github.com/VizierDB/vizier-scala/releases/tag/v2.1.1
Tom was on the National Pharmaceutical Congress podcast last week, talking about the importance of context and #Provenance, and of keeping humans in the driver's seat when using AI to support pharmaceutical research.
Grats to #BreadcrumbAnalytics on their work supporting a recent landmark study by #UBuffalo's Andy Talal on the efficacy of telemedicine in hepatitis-C treatment.
Quoth Tom:
> "By addressing the complex challenges of data reproducibility and governance, we ensure that critical research like this can lead to tangible improvements in patient care and treatment strategies."
Study Overview: https://medicine.buffalo.edu/news_and_events/news/2024/04/jama-talal-hcv-telemedicine-19662.html
Breadcrumb's Post: https://www.linkedin.com/posts/breadcrumb-analytics_ai-datascience-healthcareinnovation-activity-7264989620046577665-kZ1I
People with opioid use disorder who have hepatitis C virus were twice as likely to be successfully treated and cured if they received facilitated telemedicine treatment at their opioid treatment program.
I'm getting back into #VizierDB development after a lengthy hiatus with an experiment in polyglot IDEs. Although the experiment was not (yet) successful, it's opened up several ideas for Vizier, including ways to improve Vizier's state model, and decouple it from #ApacheSpark to also allow lighter-weight SQL engines like #DuckDB. I'm also inspired to explore #Curses as an alternative frontend to Vizier.
For now, just some maintenance with Vizier's plugin architecture.
https://github.com/VizierDB/vizier-scala/issues/288
This is now the fourth time that I've handled TA assignments for my department. It's always a pain juggling about 2-3 spreadsheets (with different views of the data) and a half-dozen websites (with supporting data).
Last year, I finally got https://vizierdb.info spreadsheet mode working, and this semester I was ready to give it a go. Holy hell was the experience far more pleasant: ~2 full days -> 4-6 hours of effort.