Automating for Consistency
summary
A custom R package called WrangleIt standardizes clinical biomarker pipelines across 45 oncology trials and 20+ external vendors, applying a single unified process driven by study-level specification files like the Data Structure file and the QC Rules file (CSV/TSV/YAML), with a Python extractor for vendor tables, GitHub-stored specs, one-line execution, scheduled cloud runs, and consistent failure/duplicate reports that replace hours of manual QC and enable faster, auditable delivery of usable data.
AI-written summary · about
chapters (11)
- 0:00Automating for consistency
- 1:28Consistent data in pharma
- 4:39Inconsistent data challenges
- 6:06Adaptable automation design
- 8:58Specification files for flexibility
- 11:50Standardizing QC rules
- 13:38Simplifying automation code
- 14:24Driving system adoption
- 16:34Automation at scale
- 17:38Key takeaways
- 17:59Q&A
related talks
11mTracking Genome Sequence Data of Cancer Patients with Docker
49mWorking on Cancer
17mOpen Source for Data Quality
11mEnsuring Reliability in Healthcare Data Engineering
41mFrictionless Compliant Data & Analytics Environments
24mProgramming Opportunities for the Global Biodiversity Observation Network
12mPackage Management for Data Scientists
35mFrom cells to drug responses - machine learning in cancer research