Updates & Roadmap

Recent updates and upcoming roadmap for Metaforms Data Validation with Python.

What's shipped recently in Data Validation with Python, and what's next on the roadmap.

For how to use the module today, see Python Data Validation. For supported question types and checks, see Validation coverage.

---

Updates

May 2026 — Cloud execution, L3 accuracy & quality checks

• Run Python validation scripts in the cloud — upload a .SAV file and execute checks directly in Metaforms without a local runtime.

• Accuracy coverage through L3 questionnaire scenarios (including ranking, multi-select filters, and related routing cases).

• Project-level quality checks:

  • Status (complete / terminate / screen-out)

  • Duplicate IDs and suspicious IP patterns

  • Straight-liners (per question, diagonal, and multi-question)

  • Trap / attention questions

  • Speeders based on length of interview (LOI)

• Config options for a unique ID column and respondent filters applied across checks.

May 2026 — Frequency tables

• Simple frequency tables and cross-tabs, with optional respondent filters.

• Descriptive statistics on single-variable frequencies (mean, median, mode, range, min, max).

June 2026 — Larger file handling

• Paginated respondent preview and MotherDuck-backed storage so larger datasets load and browse without pulling the full file into memory.

June–July 2026 — Data update & manipulation

• Modify Data to edit respondent datasets in place:

  • Rescale existing scales

  • Update a question across all respondents (with optional filters)

  • Recode values and metadata together

  • Edit string variables, run arithmetic on numerics, assign missing values

  • Create new variables and assign values (shown in Respondents; not included in SAV export)

  • Undo applied modifications

• Variable View with searchable datamap and value-label editing.

• Export a cleaned .SAV after edits.

• AI-assisted data modifications (preview and apply), plus advanced filters (computed/temp variables, conditional functions).

July 2026 — Frequency export & wide datasets

• Export frequency and crosstab tables from DV Python.

• Support for very wide surveys (on the order of ~50k columns), including streaming SAV ingest and export without loading the full file into memory.

---

Roadmap

• Data Quality — Respondent-level quality checks for both close-ended and open-ended questions, so keep/remove decisions can be made from a single, unified view.

• Change management — Carry validation forward when the questionnaire or data file changes, instead of rebuilding checks from scratch.

• L4 & advanced coverage — Scenarios and checks not yet supported end-to-end:

  • Sense checks (cross-field plausibility)

  • Frequency and quota validation against the questionnaire design

  • Hidden quota variables and hidden filter-text variables

  • Price-range plausibility checks

  • Multi-country Surveys checks