About Whitespace
What each stage of an assessment produces, and the evidence principles every figure follows. Opening a screen requires an approved account.
How it works
The walkthrough takes one disease from the question you are asking, through the evidence found and checked, to scored opportunities, a written report and a set of things to keep watching — ending with questions you can ask about that evidence. Six more screens let you dig into the underlying records whenever a number needs checking.
The wizard · 8 steps
What this run was asked to look at
The run's name, mode and model, the diseases in scope, and the weight given to each scoring dimension.
Step 2Companies and drugs the run found
Located companies grouped by disease, each carrying the source that surfaced it and the drug it relates to.
Step 3How solid the evidence base is
The run's own source-validity report, with its coverage verdict quoted verbatim from the report text.
Step 4The scoring rubric in force
Commercial, clinical and strategic frameworks with each dimension's weight and its 1/5/9 score anchors.
Step 5Composite scores and the memo behind them
Per-framework composite scores, expandable per-dimension cards with confidence and rationale, and the opportunities memo.
Step 6The deliverable, read in full
The run's strategic insights report rendered complete, with a contents list built from its own headings.
Step 7What stays watched after the run
Watch terms, monitored sources, schedule and creation time for the run's ongoing monitoring configuration.
Step 8Questions over the run's own evidence
Ask about the gathered evidence in conversation. Each thread keeps its own memory and answers cite their sources inline.
Scope of the selected run, at a glance
Trial counts for the active run_id plus phase, status, sponsor and condition distributions derived from the run's trial records.
Run-scopedEvery trial the run pulled, field by field
Search and filter the run's trials, then open a record to read its conditions, interventions, phase, status and sponsor exactly as the service returns them.
Store-wideThe asset landscape behind the trials
Normalised assets across runs, with modality, target and sponsor context for orienting a whitespace read.
Store-wideWhich assets are being combined
Co-occurring asset pairs with their observed counts, so crowded and thin combination space separate quickly.
Store-wideEntity resolution you can inspect
Candidate matches with confidence values surfaced for review. This phase is read-only: no confirm or reject writes.
Store-wideTreatment guidelines held in the store
The guideline entries behind the landscape read, listed as the service returns them.
Data principles
Every figure on every screen is fetched from the service at view time. Nothing is cached, seeded or synthesised in the frontend.
Tables and detail panels show the API's own field names next to their values, so what you read here matches what the service returned.
Where the API does not expose something yet — arm-level detail, a run index — the console says so rather than filling the space with an estimate.
Try an interactive preview, or sign in with an approved account.