About

The name reads as hi-vis. Making research visible, and inspecting it for hazards.

Most figures in the life sciences are made late, in the wrong software, by someone who has already done the hard part. The science is usually fine. The picture is usually the bottleneck. This site takes published figures, goes back to the data underneath them where the data exist, and draws them again.

The one rule

Improve the picture, never the science. A redraw may change the encoding, the colour, the layout, the annotation, the level of aggregation shown and the framing. It may not change what the data say. If a redraw makes an effect look bigger, cleaner or more significant than the original did, it is wrong and gets rebuilt.

In practice that means every piece lists every change it made, in plain English, and links to the data so you can check. Where values were read off a published image rather than obtained from a deposit, the figure says so and names the tool. Those are never presented as the authors' data.

Toward the authors

The figure is the subject. The authors are not. Assume good faith, time pressure and bad software, usually all three. Where a constraint came from the journal, such as a panel count or a single-column width, we say so, because often the journal is the culprit.

Corresponding authors are contacted before a redraw of their work is published, and offered a preview and a right of reply. Any reply is published in full, on the piece, above the fold. Corrections are permanent and collected in one place.

Counters

Figures rebuilt from data

1

Spot checks

0

Pieces with a cited accession

1

Counted from what is published, not maintained by hand.

How it is built

Charts are rendered to static SVG when the site is built, so a page carries no charting library and the figure is there whether or not scripts run. Colour comes from one palette file and nothing may name a colour of its own. The build refuses a figure without alt text, a reproduction without a recorded licence, a bar chart on a truncated baseline, and any colour that is not in the named palette.

Accessibility is checked on every build with axe in a real browser, along with reflow at 320 pixels and the keyboard behaviour of every control. The colour-vision screening of the palettes is committed to the repository rather than claimed here.

Who

Edited by Hannah Isles, PhD, previously an Associate Editor at the journal Science Immunology, now working in Sheffield. She chooses what is covered, writes the words, and signs off every piece. Hannah Isles takes on figure and manuscript work through Hannah Isles Editorial. Hannah Isles Editorial