提取真假新闻扩散网络

[chengjun]

Vosoughi, S., Roy, D., & Aral, S. (2018). The spread of true and false news online. Science, 359(6380), 1146–1151. https://doi.org/10.1126/science.aap955

import pandas as pd
import numpy as np
import json
from tqdm.auto import tqdm
import networkx as nx

raw_metadata = pd.read_csv('/Users/datalab/bigdata/FalseNews_Code_Data/data/raw_data_anon.csv')

raw_metadata

tid	veracity	cascade_id	rumor_id	rumor_category	parent_tid	tweet_date	user_account_age	user_verified	user_followers	user_followees	user_engagement	cascade_root_tid	was_retweeted
5970798468	FALSE	0	r_-1	Politics	-1	2015-05-09 16:45:03	2064	False	4353	963	2.461932	597079846846668800	0
0	FALSE	1	r_0	Politics	-1	2013-08-22 21:09:08	1326	False	164	236	7.092405	0	0
632718	FALSE	2	r_2	Politics	-1	2015-10-24 17:50:24	136	False	14149	13846	198.410726	632718	0
643829	FALSE	3	r_24	Politics	-1	2015-09-10 05:43:00	1054	False	1363	1325	3.864481	643829	0
654940	FALSE	4	r_1	Politics	-1	2013-04-22 18:46:37	1568	False	717	2071	1.428586	654940	0
...	...	...	...	...	...	...	...	...	...	...	...	...	...
624527	False	126301	r_395	Politics	624284	2016-08-30 21:58:02	2084	False	2471	2824	30.943673	613268	0
624528	False	126301	r_395	Politics	624284	2016-08-31 04:09:52	1121	False	250	805	9.176002	613268	0
624373	False	126301	r_395	Politics	624086	2016-08-31 00:46:50	49	False	8	78	0.293459	613268	0
624529	False	126301	r_395	Politics	624284	2016-08-31 20:45:42	1368	False	414	139	3.169038	613268	0
624530	False	126301	r_395	Politics	624284	2017-02-05 23:21:07	963	False	1417	709	33.920768	613268	0

4034736 rows × 14 columns

len(raw_metadata['cascade_id'].unique())

126302

from collections import Counter
c = Counter(raw_metadata['cascade_id'])
c.most_common(20)

[(82588, 46895),
(82587, 45451),
(82586, 42149),
(82585, 41196),
(82584, 38240),
(82583, 35016),
(82582, 33750),
(82581, 31515),
(126286, 31401),
(126285, 30967),
(126288, 30517),
(126287, 29229),
(82579, 26624),
(82580, 26525),
(82578, 25550),
(82577, 25033),
(82576, 24424),
(82575, 24111),
(106998, 23228),
(82574, 22330)]

c.most_common(2000)[1980:2000]

[(82838, 175),
(82855, 175),
(80924, 174),
(80928, 174),
(81103, 174),
(81119, 174),
(81220, 174),
(82857, 174),
(80835, 173),
(80920, 173),
(81002, 173),
(81013, 173),
(81108, 173),
(81212, 173),
(82797, 173),
(80826, 172),
(80827, 172),
(80906, 172),
(80907, 172),
(80911, 172)]

cascade = raw_metadata.loc[raw_metadata['cascade_id']== 80911]
len(cascade)

172

ids2date = cascade[['tid','parent_tid','tweet_date','cascade_id']]

#parent_tid=int(cascade['parent_tid'][i])
    
ids2date['tweet_date'] = pd.to_datetime(ids2date['tweet_date'])
ids2date = ids2date.sort_values(by=['tweet_date'], ascending=True)

root = ids2date[ids2date['parent_tid']== -1]
notroot = ids2date[ids2date['parent_tid']!= -1]

graph=nx.DiGraph()
graph.add_node(root['tid'].to_list()[0],date = root['tweet_date'].to_list()[0])
for j in notroot.index:
    graph.add_node(notroot['tid'][j],date = notroot['tweet_date'][j])
    graph.add_edge(notroot['parent_tid'][j],notroot['tid'][j])

import pylab as plt

plt.figure(figsize=(8, 8))

pos = nx.spring_layout(graph, k=0.8)

nx.draw(graph, pos = pos, width=0.8, 
        node_color='lightblue', node_size=20)

[chengjun]

20220812NetworkExtraction.zip