CSV is the universal currency of data work: R, Python, MATLAB, Excel, Google Sheets, and every plotting library speak it. But the data you need is often locked inside a chart image — a figure in a paper, a graph in an annual report, a plot in a slide. This guide shows how to go from graph to CSV without retyping a single number.

The three ways to get data out of a graph

1. Ask the author. The gold standard — original data beats any reconstruction. In practice, authors move on, datasets disappear, and half of "data available on request" requests go unanswered.

2. Eyeball and retype. Fine for three points on a bar chart. Hopeless for a 200-point time series, and your precision is whatever your eye guesses between gridlines.

3. Digitize the image. Calibrate the chart's axes so pixel positions map back to data coordinates, then trace the curves. This is what Plot2CSV automates.

Graph to CSV in five steps

  1. Upload a PNG or JPG of the chart at plottocsv.com/app
  2. Crop — the plot region is suggested automatically
  3. Calibrate — tap origin, max X, and max Y, then enter their values (linear or log)
  4. Extract — every line and scatter series is detected automatically and labeled Y1, Y2, …
  5. Download CSV — one file, aligned columns, edits included

The exported CSV looks exactly like what you'd want to read_csv:

X,Y1,Y2
0.0,42.1,43.0
0.5,44.6,47.2
1.0,45.2,52.9

What affects accuracy

Digitized data can be remarkably close to the source — typically within a fraction of a percent on clean figures — but three things dominate the error budget:

  • Image resolution. A 1200px-wide figure beats a 300px thumbnail every time.
  • Calibration precision. Your three reference taps define the whole coordinate system; zoom in when placing them.
  • Curve thickness and overlaps. Where two curves cross, verify the traced points in the editor — Plot2CSV keeps each series' overlay visible so mix-ups are easy to spot and drag back into place.

For scientific figures, always review the overlay before exporting — a habit that catches virtually every stray point. More on this in our guide to extracting data from scientific graphs.

No signup, no upload, no cost

Plot2CSV runs entirely in your browser: your images never leave your machine, there is no account wall, and CSV, Excel, and JSON export are all free. Convert your first graph now, or see real worked examples of multi-series extractions with the exact CSV they produce.