plotly

Plotly Best Practices

Plotly 6.5.0 is a powerful tool for interactive, publication-quality visualizations, crucial for modern AI/ML workflows. These rules ensure your Plotly code is maintainable, reproducible, and production-ready, leveraging the latest features including AI-assisted Studio tools.

Code Organization and Structure

1. Separate Data Preparation from Visualization

Keep data preprocessing, model outputs, and metric calculations distinct from plotting logic. This improves debugging and reusability.

❌ BAD:

import pandas as pd
import plotly.express as px

def analyze_and_plot_data(raw_data: pd.DataFrame):
    # Data processing mixed with plotting
    processed_df = raw_data[raw_data['value'] > 0]
    processed_df['normalized'] = processed_df['value'] / processed_df['value'].max()
    fig = px.scatter(processed_df, x='timestamp', y='normalized', title='Normalized Data')
    fig.show()

✅ GOOD:

import pandas as pd
import plotly.express as px
import plotly.graph_objects as go

def preprocess_data(raw_data: pd.DataFrame) -> pd.DataFrame:
    """Processes raw data for visualization."""
    processed_df = raw_data[raw_data['value'] > 0]
    processed_df['normalized'] = processed_df['value'] / processed_df['value'].max()
    return processed_df

def create_normalized_scatter_plot(data: pd.DataFrame) -> go.Figure:
    """Creates a scatter plot from processed data."""
    fig = px.scatter(data, x='timestamp', y='normalized', title='Normalized Data Over Time')
    return fig

# Usage example:
# raw_df = pd.read_csv('my_data.csv')
# clean_df = preprocess_data(raw_df)
# plot_fig = create_normalized_scatter_plot(clean_df)
# plot_fig.show()

2. Modular Figure Construction

Treat the Figure object as a modular data structure. Build traces, then define the layout. This mirrors Plotly's internal design and enhances readability.

❌ BAD:

import plotly.graph_objects as go

# Hard-to-read, deeply nested dictionary for figure definition
fig = go.Figure(
    data=[{'type': 'bar', 'x': [1, 2, 3], 'y': [10, 15, 13]}],
    layout={'title': {'text': 'Sales Data'}, 'xaxis': {'title': {'text': 'Month'}}}
)