Gallery

Charts

• 

  Group comparison

  Does this biomarker differ between these groups? One value across the levels of a category at chosen visits, as boxes, violins or points with the number in each group beneath, a second grouping by colour, panels by one further variable and a logarithmic scale. It opens on every biomarker at every visit, and a click on a biomarker opens it alone, with R's test of the groups under each visit. Click a box or a point to list its participants and open a participant's profile.

  Live demo · Evidence · API reference

• 

  Association scatter

  Do these two variables move together? One point per participant, with either axis a biomarker at a visit or a participant-level number, coloured by a group, in panels by one further variable, each axis linear or logarithmic. Under it R's Pearson or Spearman coefficient with its interval and p-value, for everyone drawn and within each colour; over it the line y = x, or R's linear fit or smooth with its band. Drag across the points to list the participants in a region, and click a point to open that participant's profile.

  Live demo · Evidence · API reference

• 

  Correlation matrix

  Which of these biomarkers, or which visits of one biomarker, are related? A grid over a set of variables: below the diagonal a mark sized and coloured by R's coefficient, above it the number, with the pair count of every cell, for Pearson or Spearman on the participants who have both values. It prints no p-value. It opens on every biomarker at the first visit, and a click on a cell opens that pair in the association scatter, with a way back.

  Live demo · Evidence · API reference

• 

  Biomarker screen

  Across every biomarker, where is the signal? One row per biomarker for a comparison chosen once, a standardised difference between two groups or a correlation with one variable, with R's estimate and its interval on one axis without units, and R's p-values beside it, unadjusted and adjusted across the rows by Benjamini-Hochberg or Holm. The rows are sorted by the estimate, and a click on a row opens that biomarker in the group comparison or the association scatter, with a way back.

  Live demo · Evidence · API reference

• 

  Cross-tabulation

  Is this category associated with that one? A two-way table of counts with its row and column totals and percentages, beside stacked bars of the same numbers, and R's chi-square or Fisher's exact test of it, with R's warning when an expected count is too small for chi-square. Either variable is a column or a biomarker cut at its median, tertiles, quartiles or typed points. Click a count to list its participants and open a participant's profile.

  Live demo · Evidence · API reference

• 

  Stratified survival

  Do participants with high and low levels of this biomarker have different outcomes? Kaplan–Meier curves per group, with censor marks and an at-risk strip, above a histogram of the biomarker showing where the cut falls and how many land each side. The groups are a column's or a biomarker cut at its median, tertiles, quartiles or typed points, and the cut line can be dragged: the curves follow at once, and R is asked again when it is let go. Under it R's log-rank test, each group's median survival with its interval and, for two groups, the hazard ratio with its interval. Experimental: the curves are safety.viz's estimator, which awaits its clinical review. Click a curve or a count at risk to list its participants and open a participant's profile.

  Live demo · Evidence · API reference

Shared parts

What every chart is built on. Each has the same two pages a chart has.

• Library core

  The entry point and the two committed bundles, each reporting the package version, and the two steps every chart shares before anything is drawn: a variable named one way, and named variables resolved to one row per participant with what was left out counted.

  Evidence · API reference

• Connection to R

  The one call a chart makes to get a statistic from R, answered from results shipped with the page or by R started in the browser on first use, and the shared rule for printing a p-value with its method and counts.

  Evidence · API reference

• Getting results out

  What every chart does so that what it shows can leave the browser and still say what it is: a title, a subtitle and footnotes written with placeholders filled from the view drawn, as text and never evaluated, and one footnote every chart writes last, with the date, the bio.viz version, and R's method and counts behind each statistic printed, naming the R and gsm.bio versions of a stored result; under the chart, a PNG of it with its title and footnotes drawn in, the statistics R returned and the table it drew from as CSV; and a specification, the chart's settings and filters as JSON data, that rebuilds the same chart.

  Evidence · API reference

Demo data

Every demo and most tests run on one made-up study, generated in gsm.bio from a seeded model with known effects planted in it, so a test can assert an answer that is known in advance. No real participant is in it.

• Results

  One row per participant, biomarker and visit: 11,472 rows.

  synthetic_results.csv · columns USUBJID, VISIT, VISITNUM, TEST, STRESU, STRESN

• Participants

  One row per participant: 200 rows.

  synthetic_participants.csv · columns USUBJID, ARM, SEX, AGE, BMIBL, RESPONSE

• Outcomes

  One row per participant and endpoint: 200 rows.

  synthetic_outcomes.csv · columns USUBJID, PARAMCD, PARAM, AVAL, CNSR

• Copied byte for byte from gsm.bio at commit 20e4a03, licence Apache License (>= 2). Nothing is retyped or regenerated here.
• The source record holds each file's checksum, and the unit tests fail when a file no longer matches it.