segmentation of data

[lchenbb]

In our current histogram, the data are segmented using common pandas.cut function with specific number of bins. However, this segmentation scheme tends to be useless when trying to get the positive/negative percentage. This drawback results from the nature of our data: dense around the origin while having few points far away from origin. The naive segmentation will thus create a bin around the origin (say (-100, 100)) containing a lot of data. Perhaps we need to plot the positive part and negative part separately.