How can I change the colours in a bar chart created from a data frame?

📊 Plotly Python
1

I am using this code to generate a bar plot.

import pandas as pd
import plotly.express as px

df = pd.read_csv(...)


# Melt the DataFrame to long format
melted_df = df.melt(id_vars=["Environment", "Dataset"], value_vars=["Alpha0.5Diff", "Alpha1Diff", "Alpha2Diff"],
                    var_name="α")

melted_df['α'] = melted_df['α'].replace('Alpha0.5Diff', "0.5")
melted_df['α'] = melted_df['α'].replace('Alpha1Diff', "1")
melted_df['α'] = melted_df['α'].replace('Alpha2Diff', "2")

# Create a bar chart with subplots for each "Alpha" value
fig = px.bar(melted_df, x="Environment", y="value", color="Dataset", facet_col="α",
             category_orders={"α": ["0.5", "1", "2"]},
             labels={"Environment": "Environment", "value": "Percent Difference"})

fig.update_layout(
    yaxis=dict(title_font=dict(size=24), tickfont=dict(size=24)),  # Increase y-axis label font size
    font=dict(size=20),  # Increase text font size for other elements
    # facet_col_title_font=dict(size=24),  # Increase facet (subplot) title font size
    legend=dict(font=dict(size=30)),
    barmode='overlay',
    legend_title_text="Dataset"
)

# Set opacity for the bars to distinguish overlapping bars
fig.update_traces(marker=dict(opacity=0.4), selector=dict(type='bar'))

# update the
for i in range(len(fig.data)):
    fig.update_xaxes(title_text="Environment", row=1, col=i+1, title_font=dict(size=24))

fig.update_annotations(font_size=36)  # control subplot function size

# Move the legend to the right-hand subplot at the bottom
fig.update_layout(legend=dict(x=0.89, y=0.00005))

# Show the plot
fig.write_image("...", width=2500, height=1500)

However, I would like to modify the colours used for the bars. I tried setting up a dummy column in the dataset which assigned a colour based on the dataset; however, this changes the legend to be ‘Colour: red blue green’. Is there a way to either choose colours themselves or to use the dummy column and alter the legend?

3

Hi @davidireland3 ,

If you want to change the default color sequence, you can pass a list of your custom color using keyword argument color_discrete_sequence to px.bar function.

# Create a bar chart with subplots for each "Alpha" value
fig = px.bar(melted_df, x="Environment", y="value", color="Dataset", facet_col="α",
             category_orders={"α": ["0.5", "1", "2"]},
             labels={"Environment": "Environment", "value": "Percent Difference"},
             color_discrete_sequence=[ 'red','blue','green']
)

or you can refer from the documentation.
https://plotly.com/python/discrete-color/#controlling-discrete-color-order

Hope this help.