How to efficiently plot a large number of line shapes where the points are connected two by two?

Question

I need to plot a very large number of segments with plotly. Contrary to a regular scatter plot where all points can be connected, here I need to only connect points two by two.

I considered different options:

adding line shapes to the plot; apparently relatively slow
creating a large number of line plots with only two points

Would there be a more suitable method? Possibly a single scatter plot where only every other couple of points are connected.

I'm looking for an efficient way to produce the plot in Python but also for good rendering performances.

If you separate your arrays with `None` the lines will be interrruped, e.g. `x = [1, 2, None, 3, 4]` and `y = [1, 3, None, 5, 4]`. — Maximilian Peters, Jun 23 '19 at 13:27
@MaximilianPeters Good point, this might actually be the way to go... — Cedric H., Jun 23 '19 at 20:51
Very well! I included something very similar in my code from the suggestion above, but your answer is very clear; thanks! — Cedric H., Jun 22 '20 at 08:19
@vestland [Please don’t include stand-alone tags in the title](https://meta.stackexchange.com/a/130208/1968). — Konrad Rudolph, Sep 02 '21 at 11:58
@KonradRudolph I find it practical at times, but the source says that it's not necessary so I'll take that into consideration. — vestland, Sep 02 '21 at 12:03

score 3 · Accepted Answer · answered Jun 10 '20 at 21:45

This answer builds on the suggestion in the comment from Maximilian Peters, as well as Jezraels approach to insert a new row after every nth row.

A key part is also to include fig.update_traces(connectgaps=False)

Plot:

Complete code:

import numpy as np
import pandas as pd
import plotly.express as px
import plotly.graph_objects as go

# dataframe, sample
np.random.seed(123)
cols = ['a','b','c', 'd', 'e', 'f', 'g']
X = np.random.randn(50,len(cols))  
df=pd.DataFrame(X, columns=cols)
df=df.cumsum()
df['id']=df.index

# dataframe with every nth row containing np.nan
df2 = (df.iloc[1::2]
         .assign(id = lambda x: x['id'] + 1, c = np.nan)
         .rename(lambda x: x + .5))
df1 = pd.concat([df, df2], sort=False).sort_index().reset_index(drop=True)
df1.loc[df1.isnull().any(axis=1), :] = np.nan
df1


# plotly figure
colors = px.colors.qualitative.Plotly
fig = go.Figure()

for i, col in enumerate(df1.columns[:-1]):
    fig.add_traces(go.Scatter(x=df1.index, y=df1[col],
                              mode='lines+markers', line=dict(color=colors[i])))

fig.update_traces(connectgaps=False)

fig.show()

How to efficiently plot a large number of line shapes where the points are connected two by two?

1 Answers1