/
temporal_stats.py
85 lines (68 loc) · 4.12 KB
/
temporal_stats.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
import os.path
import numpy as np
import time, datetime, calendar
import matplotlib.path as mpath
import matplotlib.patches as mpatches
import matplotlib.pyplot as plt
from Params import Params
import numpy as np
d_duration = Params.disaster_duration
for unit_bin in ['day', 'hour']:
for graph_type in ['senti_both', 'senti_score']:
for disaster_id in Params.disaster_ids:
start_date, end_date = d_duration[disaster_id][0], d_duration[disaster_id][1]
start_date = calendar.timegm(datetime.datetime.strptime(start_date, "%m-%d-%Y %H:%M:%S").timetuple())
end_date = calendar.timegm(datetime.datetime.strptime(end_date, "%m-%d-%Y %H:%M:%S").timetuple())
# conver from UTC time to PST time
start_date, end_date = start_date - 3600*8, end_date - 3600*8
for affect in ['_affected_filtered', '_unaffected_filtered']:
for filt in ['_hash', '_classify']:
filename = disaster_id + affect + filt
file = Params.without_tweet_folder + filename + '.txt'
print file
if os.path.isfile(file):
data = np.loadtxt(file, dtype=int, delimiter='\t', usecols=(1,5)) # (time, sentiment)
# neg_rows = [i for i in range(len(data)) if data[i][1] == -1]
# pos_rows = [i for i in range(len(data)) if data[i][1] == 1]
# neg_data, pos_data, all_data = data[neg_rows][:,0], data[pos_rows][:,0], data[:,0]
all_data = data[:,0]
neg_data = np.array(filter(lambda x : x[1] == -1, data))[:,0]
pos_data = np.array(filter(lambda x : x[1] == 1, data))[:,0]
if unit_bin == 'hour':
bins = (end_date - start_date)/(3600)
elif unit_bin == 'day':
bins = (end_date - start_date) / (3600*24)
# bins = 10
neg_hist, neg_edges = np.histogram(neg_data, bins)
pos_hist, pos_edges = np.histogram(pos_data, bins)
all_hist, all_edges = np.histogram(all_data, bins)
neg_ratio = np.array(neg_hist, dtype=float)/ (all_hist + 0.001)
sentiment_score = pos_hist - neg_hist
values = sentiment_score
# plot graph
fig, ax = plt.subplots()
if graph_type == 'senti_score':
index = np.arange(len(values))
bar_width = 0.2
plt.bar(index, values, bar_width, alpha=0.4, color='g', label='Sentiment score')
ticks = [str(i + 1) for i in range(len(values))]
ax.set_xticks(index + bar_width/2)
ax.set_xticklabels(ticks)
ax.set_xlabel('Time (hour)')
ax.set_ylabel('Sentiment score')
ax.legend()
ax.set_title(filename)
if graph_type == 'senti_both':
index = np.arange(len(values))
bar_width = 0.2
rects1 = ax.bar(index, neg_hist, bar_width, alpha=0.4, color='r')
rects2 = ax.bar(index + bar_width, pos_hist, bar_width, alpha=0.4, color='y')
ticks = [str(i + 1) for i in range(len(pos_hist))]
ax.set_xticks(index + bar_width)
ax.set_xticklabels(ticks)
ax.set_xlabel('Time (' + unit_bin + ')')
ax.set_ylabel('Sentiment score')
ax.legend((rects1[0], rects2[0]), ('Negative', 'Positive'))
ax.set_title(filename)
# plt.show()
plt.savefig('./output/graph/' + graph_type + '_' + unit_bin + '_' + filename + '.png', format='png', dpi=400)