What this page does
Phase-1 Data Readiness operationalizes a transparent, expert-supervised review of feature-level missingness before modeling. This interactive dashboard summarizes, by category and feature, the proportion and count of missing entries in the submitted dataset (source file → Book_10.08.2025_Dr.Rad_Final.xlsb, 1404-08-17). A pre-specified 85% missingness threshold highlights columns that are very likely uninformative for Phase-2 and surfaces them for action.
The 85% rule follows common practice in clinical prediction pipelines: extremely sparse features tend to degrade calibration and inflate variance. Use the threshold as a screening device, not an absolute rule; clinically essential variables may still be retained after review. For background see Steyerberg (2019), van Buuren (2018), and Kuhn & Johnson (2013). DOIs: 10.1007/978-3-030-16399-0 • 10.1201/9780429492259 • 10.1007/978-1-4614-6849-3
For each feature, choose Keep, Needs discussion, or Drop, and (optionally) add a brief note. Saving decisions generates a CSV (local download) and stores a copy on the server as the signed-off Phase-2 retention list. For accountability, reviewers must enter full name and confirm they are working under the supervision of Dr. Mohammad Sadighi Gilani or Dr. Marjan Sabbaghian; your confirmations here constitute the formal record.
| # | Decision | Feature | Missing (n) | Missing (%) | Suggestion | Notes |
|---|

