Breakout group on visualising networks with plotly

tom_brown · 12 October 2017 14:25

Use plotly to generate nice interactive graphics:

to visualise network flows and properties (i.e. node and edge properties of a directed graph, possibly also time series).

Minimal working example for Jupyter notebook:

gist.github.com

https://gist.github.com/nworbmot/643576828a8735d16826f7ae60fffd29

networkx_plotly.py

import plotly.offline as py
from plotly.graph_objs import *

py.init_notebook_mode(connected=True)
%matplotlib inline

import networkx as nx

G=nx.random_geometric_graph(200,0.125)
pos=nx.get_node_attributes(G,'pos')

This file has been truncated. show original

Minimal example with plot of Germany network from PyPSA using SciGRID data:

gist.github.com

https://gist.github.com/nworbmot/301099c67a87251a4431ff85f4d9c486

pypsa_plotly.py


## Based on https://plot.ly/python/network-graphs/

## Copyright plotly team, Tom Brown, Bryn Pickering 2017

## Licensed under the Apache 2.0 Licence.

## Working demo: https://pypsa.org/examples/plotly_network.html

import plotly.offline as py

This file has been truncated. show original

Rivus has some in-built plotting with plotly:

github.com

tum-ens/rivus/blob/master/rivus/io/plot.py

from pandas import Series
from numpy import union1d
import math
from mpl_toolkits.basemap import Basemap

from rivus.main.rivus import get_constants, get_timeseries, line_length
from rivus.utils.pandashp import total_bounds

COLORS = {
    # (R,G,B) tuples with range (0-255)
    # defaults
    'base': (192, 192, 192),
    'building': (192, 192, 192),
    'decoration': (128, 128, 128),
    # commodities
    'Heat': (230, 112, 36),
    'Cool': (0, 0, 255),
    'Elec': (255, 170, 0),
    'Demand': (0, 255, 0),
    'Gas': (128, 64, 0),

This file has been truncated. show original

Bad examples with static matplotlib:

brynpickering · 12 October 2017 14:31

You can plot scatters onto a map here: https://plot.ly/python/scattermapbox/

aknaut · 12 October 2017 15:06

here is an example with directed arrows:

gist.github.com

https://gist.github.com/anonymous/31e23e446f4415357aef3bbdc01d45a1

plotly_with_arrow.py

import math
import plotly.offline as py
from plotly.graph_objs import *

py.init_notebook_mode(connected=True)
%matplotlib inline

import networkx as nx

G=nx.random_geometric_graph(200,0.125)

This file has been truncated. show original

stefan.pfenninger · 12 October 2017 15:11

Here is a forked version of @tom_brown’s PyPSA example using Mapbox:

gist.github.com

https://gist.github.com/sjpfenninger/6543ebc226648429b4876405238de15c

pypsa_plotly.py

mapbox_access_token = 'PASTE_YOUR_TOKEN_HERE'

import plotly.offline as py
from plotly.graph_objs import *
import plotly.graph_objs as go

import pypsa

py.init_notebook_mode(connected=True)
%matplotlib inline

This file has been truncated. show original

aknaut · 12 October 2017 15:15

From rivus:
Greatly inspired by
Example1 <https://plot.ly/python/lines-on-maps/>_ and
Example2 <https://plot.ly/python/3d-network-graph/>_.

brynpickering · 12 October 2017 15:40

With hover information on the lines:

gist.github.com

https://gist.github.com/brynpickering/b3257ab711c27ad878bca8b1cc8ba5f1

build_network_with_hover.py

# direct copy from https://stackoverflow.com/questions/46037897/line-hover-text-in-plotly with some lines changed (originals in comments)
import plotly.offline as py
from plotly.graph_objs import *

py.init_notebook_mode(connected=True)
import networkx as nx

G=nx.random_geometric_graph(200,0.125)
pos=nx.get_node_attributes(G,'pos')

This file has been truncated. show original

tom_brown · 12 October 2017 15:51

Final demonstration of SciGRID with plotly:

https://pypsa.org/examples/plotly_network.html