Skip to content
LifeFromData

Collection possibilities

Start with an example.
Make it your study.

From DNA and bloodwork to a meal photo, a voice note or a night of sleep. Combine clinical tests, device signals and everyday logs around your research question.

Illustrative study designs, not available inventory. Source access, integrations and clinical collection are scoped for each study.

Inputs to scope

Bloodwork · Clinical records · Lab panels · Repeat tests · DNA · Genotyping · Microbiome · Molecular assays · MRI · CT · X-ray · Ultrasound · Heart rate · HRV · Sleep · Activity · Temperature · Weight · Blood pressure · Glucose · SpO₂ · Audio · Transcripts · Symptoms · Guided check-ins · Photos · Video · Movement · Visual diaries · Food · Medication · Supplements · Mood · Energy · Cycle logs

8 of 15 illustrative designs · Original illustrative photography

Example / Biomarker trajectories

Bloodwork, over time.

Connect repeat blood panels to the routines and measurements between tests.

Inputs, labels & delivery
Inputs
Source reports, analytes, units, reference intervals, collection dates
Reference labels
Source-checked observations; separate clinical interpretations
Delivery concept
JSONL or Parquet + report manifest

Example / Multimodal research

DNA meets daily life.

Scope a cohort around permitted genetic results and linked longitudinal observations.

Inputs, labels & delivery
Inputs
Genotyping or variant exports, assay metadata, approved follow-ups
Reference labels
Source variant calls and separately reviewed annotations
Delivery concept
Permitted source exports + structured metadata

Example / Image-report alignment

The scan and the story.

Link an MRI or other imaging study to its report and relevant follow-up context.

Inputs, labels & delivery
Inputs
Image series, acquisition metadata, source reports
Reference labels
Report-linked fields; specialist labels if commissioned
Delivery concept
DICOM + reports + annotation manifest

Example / Time-series modeling

Heart rate. HRV. Context.

Pair wearable heart-rate and HRV observations with activity, rest and reported events.

Inputs, labels & delivery
Inputs
Source-defined HR and HRV metrics, device metadata, coverage, event logs
Reference labels
Measured or vendor-derived metrics with method and window retained
Delivery concept
Parquet + device and coverage manifest

Example / Language understanding

Symptoms, in their own words.

Capture guided narratives and align them to a structured symptom timeline.

Inputs, labels & delivery
Inputs
Audio, timed transcripts, prompt versions
Reference labels
Reviewed transcript and extraction labels
Delivery concept
Audio + JSONL annotations

Example / Visual understanding

Movement, with context.

Collect repeat recordings of a defined movement task with capture-quality checks.

Inputs, labels & delivery
Inputs
Short videos, capture conditions, self-reports
Reference labels
Capture quality; scoped task annotations
Delivery concept
Media + annotation manifest

Example / Multimodal extraction

Food, beyond a calorie count.

Combine meal photos, spoken descriptions and timing in a repeatable food diary.

Inputs, labels & delivery
Inputs
Photographs, participant descriptions, reported portions and meal times
Reference labels
Reported ingredients; reviewed extraction; estimates marked as derived
Delivery concept
Media + JSONL event logs

Example / Multimodal modeling

Sleep, with the next morning.

Pair device sleep summaries with next-day self-reported energy and routines.

Inputs, labels & delivery
Inputs
Device summaries, coverage, sleep diaries and daily check-ins
Reference labels
Vendor-estimated sleep metrics and separate reported ratings
Delivery concept
Parquet + data dictionary

Including temperature, weight, medication, supplements and repeat photographs.

What does your model need to learn?

Start with the task. Define the collection together.

Scope a pilot