jinhangjiang

import networkx as nx

# "edgelist" will be prepared with 3 columns: Source, Target, and Weights
# "name" give your graph a name since we are going to do graphs comparison
def convert_graph(edgelist,name):
  graph=nx.Graph() # this only helps you generate a undirected graph
  graph.add_weighted_edges_from([tuple(x) for x in edgelist.values]) # add weights to the edges
  graph.name = "Network Graph for" + " " + name # name the graph
  print(nx.info(graph)) # this will give you the basic info about the graph
  print("------------------------------------")

jinhangjiang

def drawgraph(graph, density=False):
  ## insert the graph you just converted from the edge list
  ## the default for the density is False. If density = True, it will return the density of the graph
  edges,weights = zip(*nx.get_edge_attributes(graph,'weight').items()) ## extract the edges and weights for later use
  ## k will affect the optimal distance between nodes.
  ## more types of layouts can be found at https://networkx.org/documentation/stable/reference/drawing.html
  pos=nx.spring_layout(graph,k=0.05,seed=42)
  nx.draw_networkx(graph,
                   pos,
                   with_labels=True,

jinhangjiang

def drawnodegraph(graph, nodename, info=False,weightbar=0):
  # graph will be your networkx graph
  # nodename will be the node that you want to focus on
  # the default value for weightbar is 0, if increase the bar, rare relationship will be removed. Assuming no negative weights
  temp = graph.copy(as_view=False) # make a temporary graph to avoid losing original ones
  temp.remove_edges_from((e for e, w in nx.get_edge_attributes(temp,'weight').items() if w <= weightbar)) # remove rare relationhsip if weightbar is not 0
  nodelist = list(temp.neighbors(n=nodename)) #generate the nodes that have relationship with our target node
  nodelist.append(nodename) # add the target to the list
  Sub = temp.subgraph(nodelist) # draw subgraph
  edges,weights = zip(*nx.get_edge_attributes(Sub,'weight').items())

jinhangjiang

## Create undirected graphs
g <- graph_from_literal(1-2, 1-3, 1-7, 3-4, 2-3, 2-4, 3-5, 4-5, 4-6, 4-7, 5-6, 5-8, 6-7, 7-8)

## Create directed graphs using addition or substraction operators
dg <- graph_from_literal(JFK-+PEK, JFK-+CDG, PEK++CDG)

jinhangjiang

## check out the vertices and the order
V(g)
#output:
# + 7/7 vertices, named, from bd40644:
# [1] 1 2 3 4 5 6 7 8

vcount(g)
# [1] 8

## check out the edges and the size

jinhangjiang

## Adjacency list
adjlist <- get.adjlist(g)
adjlist[1]
# $Adam
# + 3/8 vertices, named, from 50df64b:
# [1] Judy  Bobby Sam  

## Edge list
as.data.frame(get.edgelist(g))
#       V1    V2

jinhangjiang

## remove specific nodes/vertices
h1 <- g - vertices(c("Adam","Judy"))

## generate subgraph by manual input
h2 <- graph_from_literal("Adam"-"Judy", "Adam"-"Bobby", "Adam"-"Sam", "Judy"-"Bobby","Judy"-"Frank")

## join graphs
h3 <- union(h1,h2) 
# or, h3 <- h1 + h2

jinhangjiang

library(sand)
g.lazega <- graph_from_data_frame(elist.lazega,
                                  directed="FALSE",
                                  vertices=v.attr.lazega)
vertex_attr_names(g.lazega)
# [1] "name"      "Seniority" "Status"    "Gender"    "Office"   
# [6] "Years"     "Age"       "Practice"  "School"   

plot(g.lazega)

jinhangjiang

#!pip install -U node2vec

# load packages
import pandas as pd
import networkx as nx
from node2vec import Node2Vec
from matplotlib import pyplot as plt, rc, cm
from matplotlib.pyplot import figure
import matplotlib.animation as animation
from mpl_toolkits.mplot3d import Axes3D

jinhangjiang

def ThreeDplot(model):
    
    "Creates TSNE model and plots it"
    "Get the labels and vectors from ndoe2vec mode"
    labels = []
    tokens = []

    for word in model.wv.vocab:
        tokens.append(model[word])
        labels.append(word)