- bokeh - micropymc PyMC in the browser using PyScript: Fair Coin?

(c) 2022, Credits: Thomas Wiecki


...fair coin...in your browser This webapp runs the following PyMC model in your browser: 
import pymc as pm

with pm.Model() as model:

    p = pm.Beta("p", alpha=1, beta=1)

    obs = pm.Binomial("obs", p=p, n=n, 

                      observed=k)

    

    idata = pm.sample()
This is a simple beta-binomial or coin-flip model. Enter how many coin-flips you did (N) and how many times it came up heads (K). Then press "Sample!" to run the NUTS sampler. For a full explanation, see this blog post. n= 10 k= 5 from pyodide import create_proxy @create_proxy def on_click(evt): k = Element('input_box_k').element.value if k == "": k = 5 n = Element('input_box_n').element.value if n == "": n = 10 run_model(n=n, k=k) Sampling will take a couple of seconds. The resulting posterior belief into what probabilities of obtaining heads are plausible will show up below:

Made using PyScript, PyMC, Arviz, Bokeh.

import warnings warnings.filterwarnings("ignore") import json from js import Bokeh, JSON from bokeh.embed import json_item from bokeh.plotting import figure import arviz as az az.rcParams["plot.backend"] = "bokeh" import os, sys sys.stderr = open(os.devnull, "w") import pymc as pm def run_model(n=10, k=5): with pm.Model() as model: p = pm.Beta("p", alpha=1, beta=1) obs = pm.Binomial("obs", p=p, n=n, observed=k) idata = pm.sample(chains=1, progressbar=False) p = figure(plot_width=500, plot_height=400, toolbar_location="below") az.plot_posterior(idata, var_names=["p"], show=False, ax=p) p.xaxis.axis_label = 'p' p_json = json.dumps(json_item(p, "myplot")) Bokeh.embed.embed_item(JSON.parse(p_json))