askmygenome

Observational, local, deterministic

One intervention. One wearable outcome. A report you can inspect.

askmygenome is a local-first N-of-1 experiment workbench. You export Apple Health XML, mark one baseline window and one intervention window, choose sleep duration, resting heart rate, or HRV SDNN, then get a cautious retrospective report and canonical JSON.

See the method Current product status
Input
Apple Health export.xml
Output
HTML view plus canonical JSON
Boundary
Retrospective observational analysis, not causal proof
Miniature result view

illustrative result

sleep duration
baseline intervention 8.1 h 6.8 h
Change
+0.72 h
Grade
promising
Uncertainty
95% interval shown in JSON

Example data only. Faded points stay on the timeline for context and are excluded from the estimate.

Method

Three steps, no black box

  1. Export

    Upload Apple Health export.xml. The product parses sleep duration, resting heart rate, and HRV as SDNN.

  2. Define windows

    Set one baseline window and one intervention window, then mark washout days, excluded dates, and dated confounders.

  3. Read cautiously

    Get a deterministic observational report with missingness checks, effect size, uncertainty, tradeoffs, warnings, and canonical JSON.

What it does

Built for one narrow question

Did one non-prescription intervention appear to improve one wearable outcome for one person, in the data they already have?

  • Explicit windows instead of fuzzy before and after summaries.
  • No zero-imputation.
  • Deterministic code as the authority.
  • Canonical JSON that records parameters, exclusions, versions, seed, and input hash.

Output

Readable report, inspectable record

Browser report

Summary, trend view, included and excluded nights, warnings, effect estimate, uncertainty, secondary tradeoffs, limitations, and a stronger next-experiment suggestion.

Canonical JSON

Engine and method versions, importer and schema versions, timezone policy, parameters, exclusions, sample counts, seed, input hash, and generation timestamp.

Privacy posture

Data stays on the user machine. The current product flow has no accounts, no persistence, and no network calls.

Status

Current product status

This is a working prototype with a local web page, CLI, deterministic engine, tests, and canonical JSON output. It currently focuses on Apple Health XML and three metrics.

The right next step is stronger method coverage and more prospective use, not louder claims.

Audience now
Careful self-experimenters who can read limitations
Deployment shape
Static marketing site, separate local-first product repo
Call to action
Read the method and current boundary before expecting more