Statistical Analysis Examples

Statistical testing and model fitting on plots

Box Plot with Significance Indicators

Add significance indicators to box plots:

import plotly.express as px
from mdu.plotly.stats import add_box_significance_indicator
from mdu.plotly.template import set_template

# Set custom template
set_template()

# Create box plot
df = px.data.tips()
fig = px.box(df, x='day', y='total_bill', title='Bill Amount by Day')

# show all, also the not significant 'ns'
fig = add_box_significance_indicator(fig, only_significant=False)

fig.show()

Statistical Model Fitting

Add regression lines and model summaries to scatter plots:

import plotly.express as px
from mdu.plotly.stats import add_statsmodel_fit

from mdu.plotly.template import set_template

set_template()

# Create scatter plot
df = px.data.tips()
fig = px.scatter(df, x='total_bill', y='tip', 
                 title='Tip vs Bill with Statistical Fit',
                 opacity=0.6)

# Add OLS regression line with confidence interval
fig = add_statsmodel_fit(
    fig,
    x=df["total_bill"].to_numpy(),
    y=df["tip"].to_numpy(),
)

fig.show()

ROC Curve Example

from mdu.plotly.ml import plot_roc_curve
import numpy as np

# Simulate model predictions
np.random.seed(42)
y_true = np.random.randint(0, 2, 100)
y_pred = np.random.rand(100)

# Add some signal
y_pred[y_true == 1] += 0.3
y_pred = np.clip(y_pred, 0, 1)

# Plot ROC curve with AUC
fig = plot_roc_curve(y_true, y_pred)
fig.update_layout(title="ROC Curve Example")
fig.show()

Perfect vs Random Classifier

# Perfect classifier
y_true_perfect = np.array([0, 0, 0, 1, 1, 1])
y_pred_perfect = np.array([0.1, 0.2, 0.3, 0.7, 0.8, 0.9])

fig_perfect = plot_roc_curve(y_true_perfect, y_pred_perfect)
fig_perfect.update_layout(title="Perfect Classifier (AUC=1.0)")
fig_perfect.show()

# Random classifier
y_pred_random = np.random.rand(100)
y_true_random = np.random.randint(0, 2, 100)

fig_random = plot_roc_curve(y_true_random, y_pred_random)
fig_random.update_layout(title="Random Classifier (AUC≈0.5)")
fig_random.show()

Color Utilities

from mdu.plotly.shared import hex_to_rgba

# Convert colors with various opacities
colors = ['#ff0000', '#00ff00', '#0000ff']
opacities = [0.3, 0.6, 0.9]

for color, opacity in zip(colors, opacities):
    rgba = hex_to_rgba(color, opacity)
    print(f"{color} at {opacity} opacity: {rgba}")

#ff0000 at 0.3 opacity: rgba(255, 0, 0, 0.3)
#00ff00 at 0.6 opacity: rgba(0, 255, 0, 0.6)
#0000ff at 0.9 opacity: rgba(0, 0, 255, 0.9)

Number Formatting

from mdu.plotly.shared import format_float_to_text_with_suffix

# Format various numbers
numbers = [1, 100, 1_000, 10_000, 1_000_000, 0.001, 0.000001]

for num in numbers:
    formatted = format_float_to_text_with_suffix(num)
    print(f"{num:>12} -> {formatted}")

           1 -> 1
         100 -> 100
        1000 -> 1000
       10000 -> 10k
     1000000 -> 1000k
       0.001 -> 1m
       1e-06 -> 1µ