distribution_plotting.ipynb not longer working with latest pymc version

[tkanngiesser]

---

AttributeError Traceback (most recent call last)
in
----> 1 ax = pm.Normal.dist(mu=0, sigma=2).plot()

AttributeError: 'TensorVariable' object has no attribute 'plot'

[maxtheman]

Here's a quick thing I put together for a binomial distribution, it should be easy to adapt to a continuous variable.

samples = pm.draw(pm.Binomial.dist(n=6, p=0.2), draws=1000) x = np.arange(np.min(samples), np.max(samples)+1) plt.plot(x, np.exp(pm.logp(pm.Binomial.dist(n=6, p=0.2), x)).eval(), "-o")