Option 1a:
This option is mainly Plotly-based with ipywidgets to interact:
# based on https://stackoverflow.com/a/73918907/8508004 and https://stackoverflow.com/q/75365695/8508004
import plotly.graph_objects as go
import pandas as pd
from ipywidgets import widgets
from ipywidgets import interact
df = pd.DataFrame(
{'ticker' : ['a', 'b', 'c', 'a', 'b', 'c', 'a', 'b', 'c'], 'timestamp' : [1,1,1,2,2,2,3,3,3], 'val' : [10,11,12,21,22,23, 100, 200, 300]})
tickers = df.ticker.tolist()
@interact
def read_values(
selected_ticker=widgets.Dropdown(
description="Select :", value="a", options=tickers[:3]
)
):
df_sub = df[df.ticker == selected_ticker]
fig = go.Figure(data=go.Scatter(x = df_sub.timestamp, y = df_sub.val))
go.FigureWidget(fig.to_dict()).show()
(Note that I find usually with Plotly Graph Objects, I have to choose something from the drop-down and then it works to display the plot. Just shows blank space until I pick that first letter. With JupyterLab it works right away. No blank space; it starts off showing the a plot immediately.)
Option 1b:
Pure Plotly-based option:
# based primarily on https://stackoverflow.com/q/69510612/8508004,
# combining with https://stackoverflow.com/q/75365695/8508004
# Influenced also by
# based on https://plotly.com/python/dropdowns/#relayout-dropdown and https://stackoverflow.com/q/71296687/8508004 and
# and https://stackoverflow.com/q/69510612/8508004 and
# https://community.plotly.com/t/how-to-delete-a-particular-trace-from-multiple-traces-in-a-graph-objects-figure/70203/2
import pandas as pd
import plotly.graph_objects as go
df = pd.DataFrame(
{'ticker' : ['a', 'b', 'c', 'a', 'b', 'c', 'a', 'b', 'c'], 'timestamp' : [1,1,1,2,2,2,3,3,3], 'val' : [10,11,12,21,22,23, 100, 200, 300]})
tickers = df.ticker.tolist()
# Create figure and add one scatter trace
fig = go.Figure()
dfa = df[df.ticker == 'a']
fig.add_trace(go.Scatter(
x=dfa.timestamp,
y=dfa.val,
visible=True,
#mode='markers',
marker=dict(size=12, line=dict(width=1, color='rgba(0.4, 0.4, 0.4, 0.8)')), # based on https://plotly.com/python/marker-style/
)
)
# Create selection buttons
select_buttons = []
for selection in tickers[:3]:
select_buttons.append(dict(method='update',
label=selection,
args=[{'x': [df[df.ticker == selection].timestamp], 'y': [df[df.ticker == selection].val]},
]
)
)
# Pass buttons to the updatemenus argument
fig.update_layout(updatemenus=[dict(buttons=select_buttons,x=0.18, y=1.21),])
# Update remaining layout properties
fig.update_layout(
title_text="Select Ticker:",
showlegend=False,
)
fig.show()
Option 2:
Matplotlib-based option:
#based on https://stackoverflow.com/a/74551129/8508004 and references therein, combined with https://stackoverflow.com/q/75365695/8508004
import matplotlib.pyplot as plt
import ipywidgets as widgets
import pandas as pd
df = pd.DataFrame(
{'ticker' : ['a', 'b', 'c', 'a', 'b', 'c', 'a', 'b', 'c'], 'timestamp' : [1,1,1,2,2,2,3,3,3], 'val' : [10,11,12,21,22,23, 100, 200, 300]})
tickers = df.ticker.tolist()
def display_selected_plot(x):
df_sub = df[df.ticker == x]
fig=plt.figure()
plt.plot(df_sub.timestamp, df_sub.val, '-o')
plt.grid(True) #optional grid
plt.show()
selected_ticker=widgets.Dropdown(
description="Select :", value="a", options=tickers[:3]
)
gui = widgets.interactive(display_selected_plot, x=selected_ticker) #Create our interactive graphic with the dropdown as the argument
gui
In my attempts, your referenced example code for Matplotlib doesn't work the way it says for me. I used ipywidget's interactive that I've seen works well lately without needing %matplotlib notebook which doesn't work in JupyterLab.
Try these out without installing anything on your system:
Go here and press 'launch binder'. A temporary, remote session will spin up served via MyBinder. For trying the Plotly one, select '3D scatter plots via Plotly' from the list available notebooks. When that opens, run all the examples using 'Run All'. Then at the bottom make a new cell & paste in the Plotly 1a example. (I don't get why the Plotly 1a example is being quirky in the classic Jupyter notebook interface and/or what isn't being imported without running that other code first in the classic Jupyter Notebook interface?) Or try Plotly 1b example.
The Matplotlib example will work if you paste it in the first notebook that comes up and run the code.
To switch to JupyterLab, click on the Jupyter icon above and on the left side. Then open a new notebook in the session that comes up. Paste in any of code blocks and run it.
As the session is temporary, if you make anything useful, be sure to download it to your local machine.
Trying on a remote temporary machine and opting-in to maintaining interactivity:
Either Plotly examples will be interactive where it works at what I suggest above to run. The options are in the upper left of the Plotly view pane.
The Matplotlib-based example has different steps to keep the 'interactive' features working depending if you are running it the classic Jupyter Notebook interface or JupyterLab.
For the classic notebook interface, at this time go to where I suggest above and simply add %matplotlib notebook at the top of the cell where you put the code I suggest. Then when you run it there, it will be interactive. You can click the blue button in the upper right to turn off the interactivity.
For keeping the interactive features available JupyterLab, at this time it is easier to use a different place to launch a temporary session. Go to here. Sadly, the link currently goes to a dead end and doesn't seem to build a new image right now. Fortunately, right now this offering works for launching. (<== Dang. That has stopped launching correctly now, too. I need somewhere where ipympl gets installed in build.) When that session opens, run in the notebook %pip install pandas. Let that run and then restart the kernel and make a new cell. At the top of the new cell, add %matplotlib ipympl as the first line above the Matplotlib suggested example. It should be interactive. The interactive tool strip will come up if you hover over the side edge just to the left of the plot area. (If not, try changing it to %matplotlib widget.)
