Tabular distribution

Examples

In app.py:

import dash_fcast.distributions as dist

import dash
import dash_bootstrap_components as dbc
import dash_core_components as dcc
import dash_html_components as html
import plotly.graph_objects as go
from dash.dependencies import Input, Output

app = dash.Dash(__name__, external_stylesheets=[dbc.themes.BOOTSTRAP])

app.layout = html.Div([
    html.Br(),
    dist.Table(
        id='Forecast',
        datatable={'editable': True, 'row_deletable': True},
        row_addable=True,
        smoother=True
    ),
    html.Div(id='graphs')
], className='container')

dist.Table.register_callbacks(app)

@app.callback(
    Output('graphs', 'children'),
    [Input(dist.Table.get_id('Forecast'), 'children')]
)
def update_graphs(dist_state):
    distribution = dist.Table.load(dist_state)
    pdf = go.Figure([distribution.pdf_plot(), distribution.bar_plot()])
    pdf.update_layout(transition_duration=500, title='PDF')
    cdf = go.Figure([distribution.cdf_plot()])
    cdf.update_layout(transition_duration=500, title='CDF')
    return [dcc.Graph(figure=pdf), dcc.Graph(figure=cdf)]

if __name__ == '__main__':
    app.run_server(debug=True)

Run the app with:

$ python app.py

Open your browser and navigate to http://localhost:8050/.

dash_fcast.distributions.Table

class dash_fcast.distributions.Table(id, bins=[0, 0.25, 0.5, 0.75, 1], prob=[0.25, 0.25, 0.25, 0.25], editable_cols=['bin-start', 'bin-end', 'pdf', 'cdf'], datatable= {}, row_addable=False, scalable=False, smoother=False, *args, **kwargs) [source]

Tabular distribution elicitation.

Parameters and attributes:

id : str, default

Distribution identifier.

bins : list of scalars, default=[0, .25, .5, .75, 1]

List of 'break points' for the bins. The first bin starts at bins[0]. The last bin ends at bins[-1].

prob : list of scalars, default=[.25, .25, .25, .25]

Probability density function. This is the amount of probability mass in each bin. Must sum to 1 and len(prob) must be len(bins)-1.

datatable : dict, default={}

Keyword arguments for the datatable associated with the table distribution. See https://dash.plotly.com/datatable.

row_addable : bool, default=False

Indicates whether the forecaster can add rows.

scalable : bool, default=False

Provides a scaling function for the table bins.

smoother : bool, default=False

Indicates whether to use a smoother for interpolation. See https://dsbowen.github.io/smoother/.

*args, **kwargs :

Arguments and keyword arguments passed to super().__init__.

Methods

elicitation(self, bins=[0, 0.25, 0.5, 0.75, 1], prob=[0.25, 0.25, 0.25, 0.25], editable_cols=['bin-start', 'bin-end', 'pdf', 'cdf'], datatable= {}, row_addable=False, scalable=False) [source]

Parameters:

bins : list of scalars or numpy.array, default=[0, .25, .5, .75, 1]

prob : list of scalars or numpy.array, default=[.25, .25, .25, .25]

datatable : dict, default={}

row_addable : bool, default=False

scalable : bool, default=False

Returns:

elicitation elements : list of dash elements

Dash elements used to elicit the distribution.

get_columns(self, editable_cols=['bin-start', 'bin-end', 'pdf', 'cdf']) [source]

Returns:	columns : list of dict List of dictionaries specifying the datatable columns. See https://dash.plotly.com/datatable,

get_data(self, bins=None, prob=None) [source]

Parameters:

bins : list of scalars or numpy.array or None, default=None

If None, use self.bins.

prob : list of scalars or numpy.array or None, default=None

If None, use self.prob.

Returns:

records : list of dict

Datatable data in records format.

register_callbacks(cls, app) [source]

Register dash callbacks for table distributions.

Parameters:	app : dash.Dash App with which to register callbacks.

fit(self, bins=None, prob=None, derivative=2) [source]

Fit the smoother given masses constraints.

Parameters:

bins : list of scalars or numpy.array

Ordered list of bin break points. If None, use self.bins.

prob : list of scalars or numpy.array

Probability density function. This is the amount of probability mass in each bin. Must sum to 1 and len(prob) should be len(bins)-1. If None, use self.prob.

derivative : int, default=2

Deriviate of the derivative smoothing function to maximize. e.g. 2 means the smoother will minimize the mean squaure second derivative.

Returns:

self :

dump(self) [source]

Dump the table distribution state dictionary in JSON format.

Returns:	state : dict, JSON

load(cls, state_dict) [source]

Load a table distribution from its state dictionary.

Parameters:	state_dict : dict Output of `Table.dump`.
Returns:	table : `Table`

pdf(self, x) [source]

cdf(self, x) [source]

pdf_plot(self, **kwargs) [source]

Parameters:	**kwargs : Keyword arguments for `go.Scatter`.
Returns:	scatter : go.Scatter. Scatter plot of the pdf.

cdf_plot(self, **kwargs) [source]

Parameters:	**kwargs : Keyword arguments for `go.Scatter`.
Returns:	scatter : go.Scatter Scatter plot of the cdf.

bar_plot(self, **kwargs) [source]

Parameters:	**kwargs : Keyword arguments for `go.Bar`.
Returns:	bar plot : go.Bar Bar plot of the pdf in the datatable.