HoloViews Introduction

Cloistered Monkey

2019-02-02 14:15

Introduction

I've already taken an initial look at HVPlot so I'm going to look at HoloViews which acts as an intermediate layer between the main plotting libraries like bokeh and matplotlib and the upper layer given by HVPlot. I haven't used it before so I'm not really sure when you would use what. I guess HVPlot gives you access to the pandas plots in bokeh without a lot of work, which is nice, although I noticed that the plots tended to be missing things sometimes (like the Hover tool) so if you want to add more back in you probably have to understand HoloViews which itself sometimes doesn't give you what you want (like the ability to render it in org-mode posts) so you still need bokeh too, sometimes. And of course I'm only using the static-page versions of everything, not the features that work with a bokeh or jupyter server. But I guess that's for later.

I'm going to be working from the Introduction of their Getting Started guide.

Set Up

Imports

Python

from functools import partial
from pathlib import Path

From PiPy

from holoviews import opts
from sklearn.datasets import fetch_california_housing
import holoviews
import numpy
import pandas

This Project

from bartleby_the_penguin.tangles.embed_bokeh import EmbedBokeh

The HoloViews Backend

If you use HVPlot you don't need to set the backend (because it defaults to 'bokeh', I think) but this is going to be about HoloViews so I'm going to do it their way, rather than relying on all the pandas methods.

holoviews.extension("bokeh")

A Partial Bokeh Embedder

Since the output folder is always the same I'm going to bind it to the EmbedBokeh definition.

plot_path = Path("../../files/posts/libraries/holoviews-introduction/")
Embed = partial(EmbedBokeh, folder_path=plot_path)

The Data Set

Load It

Sklearn downloads it as a 'bunch' so we need to get it in that form first and then turn it into a data frame (I'm sure there's a way to skip this step but this is the way I already know how to do it).

folder = Path("~/data/datasets/california-housing").expanduser()
assert folder.is_dir()
print(folder)
bunch = fetch_california_housing(folder)
print(bunch.DESCR)

/home/hades/data/datasets/california-housing
.. _california_housing_dataset:

California Housing dataset
--------------------------

**Data Set Characteristics:**

    :Number of Instances: 20640

    :Number of Attributes: 8 numeric, predictive attributes and the target

    :Attribute Information:
        - MedInc        median income in block
        - HouseAge      median house age in block
        - AveRooms      average number of rooms
        - AveBedrms     average number of bedrooms
        - Population    block population
        - AveOccup      average house occupancy
        - Latitude      house block latitude
        - Longitude     house block longitude

    :Missing Attribute Values: None

This dataset was obtained from the StatLib repository.
http://lib.stat.cmu.edu/datasets/

The target variable is the median house value for California districts.

This dataset was derived from the 1990 U.S. census, using one row per census
block group. A block group is the smallest geographical unit for which the U.S.
Census Bureau publishes sample data (a block group typically has a population
of 600 to 3,000 people).

It can be downloaded/loaded using the
:func:`sklearn.datasets.fetch_california_housing` function.

.. topic:: References

    - Pace, R. Kelley and Ronald Barry, Sparse Spatial Autoregressions,
      Statistics and Probability Letters, 33 (1997) 291-297

Make The DataFrame

data = pandas.DataFrame(bunch.data, columns=bunch.feature_names)
data["median_value"] = bunch.target
print(data.head())

   MedInc  HouseAge  AveRooms  AveBedrms  Population  AveOccup  Latitude  \
0  8.3252      41.0  6.984127   1.023810       322.0  2.555556     37.88   
1  8.3014      21.0  6.238137   0.971880      2401.0  2.109842     37.86   
2  7.2574      52.0  8.288136   1.073446       496.0  2.802260     37.85   
3  5.6431      52.0  5.817352   1.073059       558.0  2.547945     37.85   
4  3.8462      52.0  6.281853   1.081081       565.0  2.181467     37.85   

   Longitude  median_value  
0    -122.23         4.526  
1    -122.22         3.585  
2    -122.24         3.521  
3    -122.25         3.413  
4    -122.25         3.422

A Plot

Our target is the median value of the house. Does that correlate with median income?

scatter = holoviews.Scatter(data,
                            ("MedInc", "Median Income"),
                            ("median_value", "Median Value"),
                            label="California Housing")

After setting up the basic plot we can do things to affect the appearance like setting the color or adding tools.

scatter = scatter.opts(opts.Scatter(color="red", tools=["hover"]))

Adding To the Layout

What if we want to add a distrbution to the plot? HoloViews uses the + operator to indicate that you want to append a plot to another one.

layout = scatter + holoviews.Histogram(
    numpy.histogram(data.HouseAge, bins=24), kdims=["HouseAge"])
layout = layout.opts(opts.Histogram(tools=["hover"]))

Table of Contents