# Extract data from plots in papers including error bars

I would like to know how to extract data from plots in a paper and, also, how to do that in a way that include error bars. I know there are specific softwares to do that, e.g. Plot Digitizer, which also work for scanned graphs. But I want to "digitize" inside Mathematica, so I don't need external softwares.

Furthermore, the plots in a paper can be a bit polluted. An example of a graph to digitize follows below. The desired points are the black points including error bars.

• Take a look at this: mathematica.stackexchange.com/q/3831/685
– VLC
Commented Feb 26, 2013 at 16:49
• Providing a solution for a particular family of graphs is expensive, but a general solution is priceless Commented Feb 26, 2013 at 17:41
• If you want to do that without Mathematica, you can use PlotDigitizer.com. This website has a free online app that can extract data from the graph image. Commented Jan 15, 2021 at 7:59
• I had already provided the link to PlotDigitizer in my question... The comment above is not helpful. Commented Jan 15, 2021 at 20:18

In general, the problem is that error bars are plotted using different symbolic conventions in different texts, so the automatic extraction would have to be tailored somewhat to each case. An easier approach is to embed the plot into a Graphics frame using Inset, and then use the drawing tools to capture the coordinates you need by hand.

Here is a less automated but more flexible way to get data from a plot:

dataPlot = Import["https://i.sstatic.net/hoeHR.jpg"];

Manipulate[
Graphics[{}, PlotRange -> {{10, 110}, {0, 1}}, Frame -> True,
AspectRatio -> 1/GoldenRatio,
Prolog ->
Inset[Show[dataPlot, AspectRatio -> Full],
p, {Center, Center}, {w, h}]
],
{{p, {Center, Center}}, Locator}, {w, 60, 130}, {h, .5, 2}]


Here, I can now use the Locator to drag the image around inside the frame of the Graphics. The sliders let me adjust the width and height of the plot until the frame labels coincide on both axes.

To extract the data from this, click the plus sign at the top right to reveal the menu where you can select Paste Snapshot. Do this and evaluate the result to get a snapshot of the framed plot.

Now click on the plot to highlight it, followed by . to see the coordinate tool. Prsse Cmd-c to start copying the coordinates you click. In this way you can assemble the desired data sets.

• Beautiful! I just used it very effectively. It turns out you can click and lay all the points, then Ctrl-C copies the line. Commented Nov 5, 2016 at 16:49