Visualization Backends

Version

This feature was introduced in v0.7.0.

LavenderTown supports multiple visualization backends for rendering charts. By default, Altair is used, but Plotly can be enabled for interactive visualizations.

Backend Selection

The visualization backend can be selected in the charts UI component. Available backends:

• Altair (default): Fast, static charts with good performance
• Plotly: Interactive charts with zoom, pan, and 3D visualizations

Using Plotly Backend

Installation

pip install lavendertown[plotly]

Features

Plotly backend provides:

• Interactive time-series charts: Zoom and pan through time-series data
• 3D outlier visualization: Multi-dimensional scatter plots for outlier analysis
• Enhanced bar charts: Interactive ghost type distribution charts
• Heatmaps: Correlation analysis visualizations

Backend API

from lavendertown.ui.visualizations.base import get_backend

# Get Altair backend (default)
altair_backend = get_backend("altair")

# Get Plotly backend
plotly_backend = get_backend("plotly")

# Check if backend is available
if plotly_backend.is_available():
    # Use Plotly for rendering
    pass

Creating Custom Charts

Plotly Charts

from lavendertown.ui.visualizations.plotly_charts import (
    create_null_chart,
    create_outlier_chart,
    create_timeseries_chart,
    create_outlier_3d_chart,
)

# Create null value chart
null_fig = create_null_chart(data, column_name)

# Create outlier chart with bounds
outlier_fig = create_outlier_chart(data, column_name, lower_bound, upper_bound)

# Create time-series chart with anomalies
ts_fig = create_timeseries_chart(data, datetime_col, value_col, anomalies)

# Create 3D outlier visualization
outlier_3d_fig = create_outlier_3d_chart(data, ["x", "y", "z"], outliers)

Backend Implementation

AltairBackend

The default backend using Altair for static charts.

from lavendertown.ui.visualizations.altair_backend import AltairBackend

backend = AltairBackend()
backend.render_chart(st, chart, "bar")

PlotlyBackend

Interactive backend using Plotly.

from lavendertown.ui.visualizations.plotly_backend import PlotlyBackend

backend = PlotlyBackend()
if backend.is_available():
    backend.render_chart(st, figure, "line")

Fallback Behavior

If Plotly is not installed, the system automatically falls back to Altair. The UI will show an error message if a requested backend is not available.