Implementation:Evidentlyai Evidently SDK Panels

Knowledge Sources	Evidentlyai_Evidently
Domains	SDK, Dashboard, Visualization
Last Updated	2026-02-14 12:00 GMT

Overview

Provides factory functions for creating dashboard panel configurations including text, counter, line plot, bar plot, and pie chart panels.

Description

The SDK Panels module offers a set of convenience factory functions that create DashboardPanelPlot instances with pre-configured plot_params for different visualization types. Each function returns a DashboardPanelPlot with the appropriate plot_type set in its parameters.

Available factory functions:

• text_panel(title, description, size) -- Creates a text-only panel for displaying headers, descriptions, or section labels. Takes no metric values. The plot_type is set to "text".

• counter_panel(title, values, description, size, aggregation) -- Creates a counter panel that displays metric values as prominent numerical counters. Supports configurable aggregation (e.g., "last", "sum", "mean") for time series data. The plot_type is set to "counter".

• line_plot_panel(title, values, description, size) -- Creates a line plot panel for visualizing metrics over time as continuous lines. The plot_type is set to "line".

• bar_plot_panel(title, values, description, size, stacked) -- Creates a bar chart panel for comparing metric values. Supports stacked bars via the is_stacked plot parameter. The plot_type is set to "bar".

• pie_plot_panel(title, values, description, size, aggregation) -- Creates a pie chart panel for showing proportional distributions. Supports configurable aggregation for time series data. The plot_type is set to "pie".

All functions accept a size parameter (default: "full") and return DashboardPanelPlot instances that can be added to DashboardModel configurations.

Usage

Use these factory functions when programmatically building dashboard configurations for Evidently projects. Combine them with PanelMetric objects from evidently.sdk.models to specify which metrics appear in each panel. Add the resulting panels to a DashboardModel for the project's dashboard.

Code Reference

Source Location

• Repository: Evidentlyai_Evidently
• File: src/evidently/sdk/panels.py

Signature

def text_panel(
    title: str,
    description: Optional[str] = None,
    size: str = "full",
) -> DashboardPanelPlot:
    ...

def counter_panel(
    title: str,
    values: List[PanelMetric],
    description: Optional[str] = None,
    size: str = "full",
    aggregation: str = "last",
) -> DashboardPanelPlot:
    ...

def line_plot_panel(
    title: str,
    values: List[PanelMetric],
    description: Optional[str] = None,
    size: str = "full",
) -> DashboardPanelPlot:
    ...

def bar_plot_panel(
    title: str,
    values: List[PanelMetric],
    description: Optional[str] = None,
    size: str = "full",
    stacked: bool = False,
) -> DashboardPanelPlot:
    ...

def pie_plot_panel(
    title: str,
    values: List[PanelMetric],
    description: Optional[str] = None,
    size: str = "full",
    aggregation: str = "last",
) -> DashboardPanelPlot:
    ...

Import

from evidently.sdk.panels import (
    text_panel,
    counter_panel,
    line_plot_panel,
    bar_plot_panel,
    pie_plot_panel,
)

I/O Contract

Inputs

Name	Type	Required	Description
title	str	Yes	Title displayed at the top of the panel
values	List[PanelMetric]	Yes (except text_panel)	List of PanelMetric objects specifying which metrics to display
description	Optional[str]	No	Optional subtitle/description for the panel
size	str	No	Panel size (default: "full"); other options include "half"
aggregation	str	No (counter_panel, pie_plot_panel only)	Aggregation method for time series data (default: "last"); options include "sum", "mean"
stacked	bool	No (bar_plot_panel only)	Whether to stack bars (default: False)

Outputs

Name	Type	Description
text_panel return	DashboardPanelPlot	Panel with plot_type="text" and empty values list
counter_panel return	DashboardPanelPlot	Panel with plot_type="counter" and specified aggregation
line_plot_panel return	DashboardPanelPlot	Panel with plot_type="line"
bar_plot_panel return	DashboardPanelPlot	Panel with plot_type="bar" and optional is_stacked parameter
pie_plot_panel return	DashboardPanelPlot	Panel with plot_type="pie" and specified aggregation

Usage Examples

from evidently.sdk.panels import text_panel, counter_panel, line_plot_panel, bar_plot_panel, pie_plot_panel
from evidently.sdk.models import PanelMetric, DashboardModel, DashboardTabModel

# Create metrics to display
drift_metric = PanelMetric(metric="DatasetDriftMetric", legend="Drift Score")
accuracy_metric = PanelMetric(metric="ColumnSummary", legend="Accuracy", metric_labels={"column": "accuracy"})

# Create a text header panel
header = text_panel(title="Model Monitoring Dashboard", description="Daily metrics overview")

# Create a counter panel showing latest drift score
drift_counter = counter_panel(
    title="Current Drift Score",
    values=[drift_metric],
    aggregation="last",
    size="half",
)

# Create a line plot showing accuracy over time
accuracy_trend = line_plot_panel(
    title="Accuracy Trend",
    values=[accuracy_metric],
    description="Daily accuracy values",
)

# Create a stacked bar chart
stacked_bars = bar_plot_panel(
    title="Metric Comparison",
    values=[drift_metric, accuracy_metric],
    stacked=True,
)

# Create a pie chart
distribution = pie_plot_panel(
    title="Label Distribution",
    values=[PanelMetric(metric="ColumnSummary", legend="Labels")],
    aggregation="last",
)

# Assemble into a dashboard
panels = [header, drift_counter, accuracy_trend, stacked_bars, distribution]
dashboard = DashboardModel(
    tabs=[DashboardTabModel(title="Overview", panels=[p.id for p in panels])],
    panels=panels,
)

Related Pages

• Environment:Evidentlyai_Evidently_Python_Core_Environment