StackedBarChart

This example, drawn from Datawrapper’s official documentation, demonstrates how to create a diverging stacked bar chart showing trust levels in media reporting across different topics. The chart displays percentages with a custom color scheme to differentiate between trust levels, sorted by low trust values. It also includes value formatting.

import pandas as pd
import datawrapper as dw

# Load media trust data from GitHub
df = pd.read_csv(
    "https://raw.githubusercontent.com/chekos/Datawrapper/main/tests/samples/stacked_bar/media-trust.csv",
    sep=";"
)

chart = dw.StackedBarChart(
    # Chart title
    title="Trust in Media Reporting",
    # Introductory text explaining the context
    intro="Trust in Media Reporting regarding widely reported topics of 2015",
    # Data source attribution
    source_name="Infratest dimap",
    source_url="http://www.infratest-dimap.de/umfragen-analysen/bundesweit/umfragen/aktuell/wenig-vertrauen-in-medienberichterstattung/",
    # Data from pandas DataFrame
    data=df,
    # Enable percentage stacking
    stack_percentages=True,
    # Format value labels as percentages
    value_label_format="0%",
    # Use diverging mode for better visual separation
    value_label_mode="diverging",
    # Sort bars by the "Low trust" column in descending order
    sort_bars=True,
    sort_by="Low trust",
    reverse_order=True,
    # Enable the color legend
    show_color_key=True,
    # Use thick bars for better visibility
    thick_bars=True,
    # Use block labels for category names
    block_labels=True,
    # Custom color mapping for each trust level
    color_category={
        "Very high trust": "#15607a",
        "High trust": "#719aae",
        "No answer": "#e8e8e8",
        "Low trust": "#ff6954",
        "Very low trust": "#c71e1d",
    },
)

# Create the chart in Datawrapper
chart.create()

Reference

Parameter	Default	Description	Type
`aria_description`	“”	The alternative text for screen readers	str
`auto_dark_mode`	False	Whether the chart should automatically flip to dark mode when the user’s system is in dark mode	bool
`base_color`	0	The base color (can be hex string or palette index)	str or int
`block_labels`	False	Whether to use block labels	bool
`byline`	“”	The byline that appears below the chart	str
`chart_id`	None	The chart ID after creation (populated by create() method)	str or None
`chart_type`	“d3-bars-stacked”	The type of datawrapper chart to create	Literal[d3-bars-stacked]
`color_category`	PydanticUndefined	A mapping of layer names to colors	dict[str, str]
`custom`	PydanticUndefined	A dictionary of custom tags to attach to the chart	dict[str, Any]
`dark_mode_invert`	True	Whether to invert colors in dark mode	bool
`data`	PydanticUndefined	The data to use for the chart	DataFrame or list[dict]
`date_label_format`	“”	The date format. Use DateFormat enum for common formats or provide custom format strings.	DateFormat (AUTO, YEAR_FULL, YEAR_TWO_DIGIT, …) or str
`download_image`	False	Whether to allow PNG download	bool
`download_pdf`	False	Whether to allow PDF download	bool
`download_svg`	False	Whether to allow SVG download	bool
`embed`	False	Whether to allow embedding	bool
`force_attribution`	False	Whether to attribute the chart to datawrapper	bool
`forkable`	True	Whether to allow other users to fork this visualization	bool
`get_the_data`	False	Whether to allow data downloads	bool
`groups_column`	None	The column to use for grouping	str or None
`hide_title`	False	Whether or not to hide the title	bool
`intro`	“”	The intro text that appears above the chart	str
`language`	“en-US”	The locale of the chart, which defines decimal and thousand separators as well as translations of month and weekday names.	str
`logo`	False	Whether to show a logo	bool
`logo_id`	“”	The id of the logo to show	str
`negative_color`	None	The negative color to use, if you want one	str or None
`notes`	“”	The footnotes that appear below the chart	str
`range_value_labels`	“”	The field you want to use for the value labels	str
`replace_flags`	“off”	Whether to replace country codes with flags. Use ReplaceFlagsType enum for type safety or provide raw strings.	ReplaceFlagsType (OFF, FOUR_BY_THREE, ONE_BY_ONE, …) or str
`reverse_order`	False	Reverse the order of the ranges	bool
`share_buttons`	False	Whether to show social media share buttons	bool
`share_url`	“”	What URL to share	str
`show_color_key`	False	Enables the legend	bool
`sort_bars`	False	Whether to sort bars	bool
`sort_by`	“”	Which column to sort by	str
`sort_ranges`	False	Whether to sort the ranges	bool
`source_name`	“”	The source name that appears below the chart	str
`source_url`	“”	The source URL that appears below the chart	str
`stack_percentages`	False	Whether to display values as percentages	bool
`theme`	“datawrapper”	The theme of the chart	str
`thick_bars`	False	Whether to use thick bars	bool
`title`	“”	The headline that appears above the chart	str
`transformations`	PydanticUndefined		Transform or dict[str, Any]
`value_label_format`	“”	The number format for value labels. Use DateFormat for temporal data, NumberFormat for numeric data, or provide custom format strings.	DateFormat (AUTO, YEAR_FULL, YEAR_TWO_DIGIT, …) or NumberFormat (AUTO, THOUSANDS_WITH_OPTIONAL_DECIMALS, INTEGER, …) or str
`value_label_mode`	“left”	How to place the over-bar labels. Use ValueLabelMode enum for type safety or provide raw strings.	ValueLabelMode (LEFT, DIVERGING) or str