Add legends by color in Plotly scatter plot

Question

trace = go.Scatter(
x=x_pca_df['Principle'],
y=x_pca_df['Second'],
mode='markers',
marker=dict(color=home_data['PriceLevel'], size=4, showscale=False))

data = [trace]

layout = dict(
title='PCA Analysis',
xaxis=dict(title='First Principle Component'),
yaxis=dict(title='Second Principle Component'))
fig = dict(data=data, layout=layout)
iplot(fig)

I want to have legends shown by the side of this plotly chart. Therefore, people can understand what the color of the points stand for. I cannot find a proper solution anywhere.

Answer 1

I think you need to specify filename parameter. Try this and let me know if it works.

trace = go.Scatter(
x=x_pca_df['Principle'],
y=x_pca_df['Second'],
mode='markers',
marker=dict(color=home_data['PriceLevel'], size=4, showscale=False))

data = [trace]

layout = dict(
title='PCA Analysis',
xaxis=dict(title='First Principle Component'),
yaxis=dict(title='Second Principle Component'), showlegend = True)
fig = dict(data=data, layout=layout)
iplot(fig, filename = 'show-legend')

Answer 2

You need to specify parameter name in your trace . As I think from seeing the plot, you have four traces . So you need to select name for each trace and choose what you want to see for each of them.

Code:

trace = go.Scatter(
x=x_pca_df['Principle'],
y=x_pca_df['Second'],
#Set parameter name to what you want to see in legend
name = 'PriceLevel',
mode='markers',
marker=dict(color=home_data['PriceLevel'], size=4, showscale=False))

data = [trace]

layout = dict(
title='PCA Analysis',
xaxis=dict(title='First Principle Component'),
# Do not forget specify showlegend as True
yaxis=dict(title='Second Principle Component'), showlegend = True)
fig = dict(data=data, layout=layout)
# Parameter filename just create a html file in your python script directory with name
iplot(fig, filename = 'show-legend')

Answer 3

I think the issue here is two fold, first your PriceLevel is an integer so the plot is using it as a scale, which is why you specified showscale=False . If you change that to showscale=True you will get a legend of sorts but it will be a scale and you want points.

That's the second part; if you want to show them as individual scales you will have to set them up as individual traces.

For Example

# Set up a trace for each level of PriceLevel
trace1 = go.Scatter(
x=x_pca_df.query(" PriceLevel==1")['Principle'],
y=x_pca_df.query(" PriceLevel==1")['Second'],
# Add a name for each trace to appear in the legend
name = 'PriceLevel 1', 
mode='markers',
marker=dict(color='rgba(152, 0, 0, .8)', size=4, showscale=False))

trace2 = go.Scatter(
x=x_pca_df.query(" PriceLevel==2")['Principle'],
y=x_pca_df.query(" PriceLevel==2")['Second'],
name = 'PriceLevel 2', 
mode='markers',
marker=dict(color='rgba(255, 182, 193, .9)', size=4, showscale=False))

# Join them all together 
data = [trace1, trace2]

layout = dict(
title='PCA Analysis',
xaxis=dict(title='First Principle Component'),
yaxis=dict(title='Second Principle Component'))
fig = dict(data=data, layout=layout)
iplot(fig)

Hopefully that'll work.