How to extract data from a line graph
A line graph is a table with its numbers taken away. This puts them back: one series for each line, one row for each point.
What comes back
- A line with markers: one row for each marker.
- A smooth curve without markers: points along the curve, close enough together to draw it again.
- Dashed and dotted lines are told apart by their dashes, and named from the legend.
- Level and upright guide lines (a zero line, a threshold) are listed apart from the data.
Steps
- Drop in the graph (the PDF if you have it).
- Look at the line drawn back over each curve.
- Correct anything in the table, then download Excel or CSV.
How exact it is
When a PDF holds the chart as a drawing (most figures in papers do), the values are read from the drawing itself. Nothing is estimated from pixels, so a point is as precise as the file is.
From pictures of charts we drew ourselves, 280 of 300 were read fully right. On figures from recent papers the values read from a picture matched the values in the PDF on 28 of 53, so check a picture reading against the overlay.
Questions
How many points do I get from a smooth curve?
As many as the curve needs: more where it bends. From a PDF they are the curve's own points, exactly as it was drawn.