how to plot (plotly.express) multiple lines based on different dataFrame parts

user_3512538 0
Viewed 2501

I have a dataFrame for example:

In [1]: import pandas as pd

In [2]: df = pd.DataFrame(data = [{'key': '1', 'value': '1', 'metadata': '3'}, {'key': '2', 'value': '2', 'metadata': '3'}, {'key': '1', 'value': '3', 'metadata': '3'}, {'key': '2', 'value': '4', 'metadata':
   ...: '3'}])

In [3]: df
Out[3]: 
  key value metadata
0   1     1        3
1   2     2        3
2   1     3        3
3   2     4        3

I want to split the df by the "key", i.e.

In [4]: df_list = [d for _, d in df.groupby(['key'])]

In [5]: df_list
Out[5]: 
[  key value metadata
 0   1     1        3
 2   1     3        3,
   key value metadata
 1   2     2        3
 3   2     4        3]

Now I have list of N (2) dataFrames based on number of unique keys. How can I plot this using plotly?

I can

In [6]: import plotly.express as px

In [7]: fig = px.line(df_list[0])

but how can I add the other lines? plotly.express.Figure has no add_line method...

plotly-express pandas plotly python
1 Answers

Running plotly.express wil return a plotly.graph_objs._figure.Figure object. The same goes for plotly.graph_objects with go.Figure(). So after building a figure using plotly express, you can add lines or traces using:

fig.add_trace(go.Scatter)

Or:

fig.add_scatter()

Other options are:

fig.add_area()
fig.add_bar()

After building a fig you can run dir(fig) to learn more.

So in a dataframe such as this:

   2021  2022  2023
0     0     0     0
1     0    -1    -2
2    -1    -1    -3
3    -1    -2    -2
4    -1    -1    -3

You can add line / trace to an existing figure using:

fig.add_trace(go.Scatter(x = df.index, y = df['2023']))

If you prefer to keep control of subsets of your data through dictionaries, one approach to add several traces is this:

extra = {'2024': df['2024'],
         '2025': df['2025']}

for k, v in extra.items():
    fig.add_scatter(x=v.index, y = v, name = k + ' from dict' )

enter image description here

Complete code:

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

# data
np.random.seed(1)
start = 2021
ncols = 6
nrows = 1000
cols = [str(i) for i in np.arange(start, start+ncols)]
df = pd.DataFrame(np.random.randint(-1,2, (nrows,ncols)), columns = cols).cumsum()
df.iloc[0] = 0

# figure
fig = px.line(df, x=df.index, y=cols[:-4])
fig.update_layout(template = 'plotly_dark')

# fig.add_trace(go.Scatter(x = df.index, y = df['2023'], name = '2023 = added trace'))

fig.add_scatter(x = df.index, y = df['2023'], name = '2023 from add_trace')


extra = {'2024': df['2024'],
         '2025': df['2025']}

for k, v in extra.items():
    fig.add_scatter(x=v.index, y = v, name = k + ' from dict' )

fig.show()
user_34377870
Related