DataThief III by Bas Tummers
Ever have the need to pull data from a publication but you cannot access the raw data. Perhaps it is from an old publication and the graph was hand drawn or the data files that went into making the graph have been long lost through the evolution of data storage media. Estimating and inputting data values by eye can be time consuming and prone to errors. Luckily there are software solutions.
One solution is DataThief, a Java script program that allows you to input an image file of a graph and gives you several tools to pull out graphing data. The program is available for free online at http://www.datathief.org/ and has a thorough user’s manual. The program can be an effective way to reproduce and re-plot graphs where the original data files cannot be procured. There are several methods to estimate secondary data values from the graphs that make it flexible for many graphics you may encounter. By establishing the origin as well as the max and min x-axis and y-axis variables, then the beginning and end point of a line graph, continuous value estimates at a user-defined interval can be established. A second method involves placing markers along the line graph and getting point estimates based on those positions. The output comes as a delimited text file that can be manipulated in spreadsheets or R. Some tips I had learned were to ensure axes of graphs that I scanned in had to be rotated slightly in a graphics programs to make them perfectly horizontal and vertical. Colour images of JPEGs had to be converted to 24-colour file formats or there was an error importing the image file.