I stumbled across an interesting error while analyzing data today. The following image is a contour generated from the result of the INT_TABULATED function in IDL. Without giving away too much of this work in progress I can say that the contour represents the position of something I am chasing. Red means there is a lot of it at the location and black means it is entirely absent.

Same result as the contour below, except this was processed without an important keyword set.

The black marks spread across the contour are not intentional. While they would represent a fascinating result if they were real, they are too blocky and bit-noisy to be authentic.

The function INT_TABULATED is used to perform integrations (i.e., finding the area under a curve) and is something that I have only recently started using. Short story shorter, if you give this function data that is not already ordered from smallest to largest, then you must also pass it the SORT keyword. The reference guide makes it clear that this is necessary for data that is in random order but my data was simply reversed, going from largest to smallest, so I did not think it necessary to reorder it.

After sorting this all out the following, accurate, result is generated.

A sensible data contour.

This demonstrates how even a minor detail can lead to a big problem and lost time. Now I better get back to investigating that red blob.

Tags: IDL, data analysis