How to plot periodic data with matplotlib

Question

My data recording system has center values [0,10,20,...350], and the data is assumed to collected evenly from the center points. Therefore, I want to plot histogram binned in this way: [-5,5],[5,15],[15,25] ... [345, 355].

The problem is that, the data are stored in [0, 360], which would make the inital and end sector just half of the data.

Simply put, when I want to plot the data for [-5,5], the actual data ploted would just have half the size , since [-5,5] is split into [0,5] and [355,360]. You can look at the 0 , 355 bin at the pic below.

bins=np.arange(-5, 360 + 10, 10)
df['dir'].hist(bins=bins, alpha=0.5, figsize=(15, 3))

My questions are:

1. What would be a good way to plot circular data, so I can solve the half size problem?
2. The plot has an emply sector start at -50 , is there way to get rid of it?

Answer 1

1: if you don't want the beautiful polar plot suggested by roadrunner, and instead want to keep your linear bar chart, you should assign any values > 355 to the 0 bin, eg,

wrapped = np.array(df['dir'])
wrapped[wrapped > 355] = 0
plt.hist(wrapped, bins=np.arange(0,370,10), align='left')

(align='left' plots your bins from, eg, -5..5, 5..15, etc given data values ranging from 0..355)

2: now get rid of that unwanted space on the left with:

plt.axis('tight')