How to hide legend with Plotly Express and Plotly

Question

I am trying to learn Plotly by firstly creating a simple bar chart in Plotly Express and then updating it with Plotly to finesse it. I would like to hide the legend.

I am trying to update the original figure by hiding the legend, and I can't get it to work. This is my traceback error.

And my code

import plotly.plotly as py
import plotly.graph_objs as go
import plotly_express as px
import pandas as pd


df = pd.read_csv('C:/Users/Documents/Python/CKANMay.csv')

fig = px.bar(df, x="Publisher", y="Views", color="Publisher", barmode="overlay")

fig.update(fig, showlegend="false")

This is what the chart looks like now with the legend. Basically, I want that awful legend on the right to go away

Answer 1

After creating the figure in plotly, to disable the legend you can make use of this command:

fig.update_layout(showlegend=False)

For Advanced users: You can also enable/disable the legend for individual traces in a figure by setting the showlegend property of each trace. For instance:

fig.add_trace(go.Scatter(
    x=[1, 2],
    y=[1, 2],
    showlegend=False))

You can view the examples here: https://plotly.com/python/legend/

Answer 2

try this:

my_data = [go.Bar( x = df.Publisher, y = df.Views)]
my_layout = go.Layout({"title": "Views by publisher",
                       "yaxis": {"title":"Views"},
                       "xaxis": {"title":"Publisher"},
                       "showlegend": False})

fig = go.Figure(data = my_data, layout = my_layout)

py.iplot(fig)

• the argument showlegend is part of layout object which you did not specify in your code
• The code may also work if you do not wrap the layout object my_layout inside a go.Layout(). It could work by simply keeping my_layout a dictionary
• as for plotly.express, try fig.update_layout(showlegend=False). All the arguments are the same. Accessing them varies slightly.

Hope it works for you.

Answer 3

The issue with your original code is that fig.update() doesn't take fig as an argument. That line could just be fig.update(layout_showlegend=False)