Different histogram under the same number of bins about practical-statistics-for-data-scientists HOT 2 CLOSED

The shape of a histograms depends on how the bins are defined. You can see in the lets-plot graph that the left bin starts at around -2.5 and the right bin ends at about 40. For the pandas plot, the bins range from 0 to about 38. While it hardly changes the overall shape, the distribution of cases into the first two bins changes with the observed reversal in the column heights.

If I compare the two graphs, I would say that the bins that lets-plot chooses are not appropriate for the data set. If you look at the lets-plot graph, the left most bin from -2.5 to 2.5, the population count is however only a positive number, so half of the bin range is not representing data.

The documentation for geom_histogram (https://lets-plot.org/pages/api/lets_plot.geom_histogram.html) shows that there are keyword arguments that allow to control the positioning of the bins. Try using for example boundary=0 and see how that changes the overall shape.

If the datasets aren't too large, I like to use density plots. They often give a more informative picture of distributions.

Examples:

• https://chartio.com/learn/charts/histogram-complete-guide/#choose-interpretable-bin-boundaries

from practical-statistics-for-data-scientists.

Hi @gedeck, thanks for replying to the issue so quickly. Oh dear, how careless I am to miss that detail :( . I will close this issue since it's just my code error.

from practical-statistics-for-data-scientists.