How to extract data from a figure with several panels
Most figures in papers are not one chart but several, side by side. Each panel has its own axes, so each has to be read by itself.
How panels are read
From a PDF, every plot area on every page is found and read against its own axes. A label shared by two panels goes to both.
From a picture, the charts in it are found first and then read one at a time.
What we measured
- From PDFs: on 200 pages holding one to six charts each, 602 of 603 charts were read correctly, and nothing that is not a chart was read as one.
- From pictures: all 30 charts in pictures of several were found, and 27 were read right.
Steps
- Upload the paper's PDF.
- Pick a chart from the tabs: panels on one page are lettered a, b, c in reading order.
- Download Excel: one sheet for each chart.
Limits
A small chart set inside another (an inset) is left unread, and the result says so: its curve belongs to its own little axes.
Small, crowded figures are the hard case for pictures: when a plot is only a couple of hundred pixels high, a pixel is already a visible error. Use the PDF when you have it, or a larger screenshot.
Questions
Can I read just one panel?
Yes. Take a screenshot of that panel and drop it in, or pick its tab after uploading the PDF.