Prepping Data for Histograms

[mtanco]

This histogram plot examples in the Wave Docs show how to display data when you already know the low and high value for each bin. Often I find that I have raw data that I would like to both bin and display as a histogram. If you're using pandas or numpy, you can use the np.histogram() function to create bins. Here's an example I regularly reference in case this is helpful to anyone else:

import numpy as np
from h2o_wave import site, data, ui

# Original data with at least one numeric column we would like to visualize distributions of
number_of_rows_of_fake_data = 100
numeric_data = np.random.randint(low=0, high=1000, size=number_of_rows_of_fake_data)

# Get bins and count for each bin using the numpy histogram function
# `bins` is an array one item larger than `counts` as it has the high and low values in one
# You can also use `np.histogram(my_pandas_dataframe['my_numeric_column'])` if you data is in pandas
counts, bins = np.histogram(numeric_data)

page = site['/demo']

page['histogram_plot'] = ui.plot_card(
    box='1 1 8 8',
    title='Count of Product Prices',
    data=data(
        fields=['count', 'low', 'high'],  # use any names here for reference in the `ui.mark` parameters
        columns=[  # all columns need to be in a list datatype, not `ndarray`
            counts.tolist(),
            bins[:(len(bins) - 1)].tolist(),
            bins[1:].tolist(),
        ],
        pack=True,  # use `pack` to take up less memory, for cases when the data in your plot is not going to change
    ),
    plot=ui.plot([ui.mark(
        type='interval',
        y='=count', y_title='Number of Products',
        x1='=low', x2='=high', x_title='Product Price',
        color='green',
        stroke_color='black',
    )])
)

page.save()