twitter_sentiment_analysis/democrats_analysis.py at master · ahwkuepper/twitter_sentiment_analysis · GitHub

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
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
#!/usr/bin/env python
# -*- coding: utf-8 -*-

#for reading and writing
import pickle

#for plotting
import matplotlib as mpl
import matplotlib.pyplot as plt

#for statistics on tweeted words
import numpy as np
import pandas as pd


"""Sentiment analysis of keyword lists using the ANEW ratings (Bradley & Lang 1999)
    """


#############
# functions #
#############

#I/O of python objects via pickle files
def save_obj(obj, name ):
    with open('obj/'+ name + '.pkl', 'wb') as f:
        pickle.dump(obj, f, pickle.HIGHEST_PROTOCOL)

def load_obj(name ):
    with open('obj/' + name + '.pkl', 'rb') as f:
        return pickle.load(f)

#get value of a 2d Gaussian distribution with mean mu's and stddev sigma's
def gaussian2d(x, y, mux, muy, sigmax, sigmay, norm):
    return np.exp(-0.5*(np.power((x - mux)/sigmax, 2.) + np.power((y-muy)/sigmay, 2.0)))/(2.0*np.pi*sigmax*sigmay)


#################
# data analysis #
#################

#read in pickle files with counter arrays
clinton_counter = load_obj("clinton_democrats_debate_03072016")
sanders_counter = load_obj("sanders_democrats_debate_03072016")

#create dictionaries from arrays
dict_clinton = {}
dict_sanders = {}

for word,count in clinton_counter:
    dict_clinton[word] = count
for word,count in sanders_counter:
    dict_sanders[word] = count


#read ANEW csv file with Pandas and create dictionary
dict_anew = {}
df = pd.read_csv('anew-1999/all.csv')
for i in np.arange(len(df)):
    dict_anew[df.loc[i,'Description']] = i


#create a 10x10 units grid for the valence and arousal ratings
xvec = np.arange(0,10,0.1)+0.5
yvec = np.arange(0,10,0.1)+0.5
x_grid, y_grid = np.meshgrid(xvec, yvec)


#select the dictionaries to use in the analysis
candidates = ['clinton', 'sanders']


#match the candidate dictionaries with the ANEW dict and sum up contributions of the words in the grids
for candidate in candidates:
    emo_grid = np.zeros(np.shape(x_grid)) #iniitalize as 0
    exec("dict_plot = dict_"+ candidate)
    for word in dict_plot.keys():
        if word in dict_anew.keys():
            valence = df.loc[dict_anew[word], 'Valence Mean']
            dvalence = df.loc[dict_anew[word], 'Valence SD']
            arousal = df.loc[dict_anew[word], 'Arousal Mean']
            darousal = df.loc[dict_anew[word], 'Arousal SD']
            dominance = df.loc[dict_anew[word], 'Dominance Mean']
            ddominance = df.loc[dict_anew[word], 'Dominance SD']
            for i in np.arange(len(x_grid)):
                emo_grid[i] = emo_grid[i] + gaussian2d(x_grid[i],y_grid[i],valence,arousal,dvalence,darousal,dict_plot[word]) #add contribution of word across the grid (normalized by occurence of key word)
    #                emo_grid[i] = emo_grid[i] + gaussian2d(x_grid[i],y_grid[i],valence,dominance,dvalence,ddominance,dict_plot[word]) #switch if dominance is wanted instead of arousal

    #normalize each grid to 1 (total sum of fields in the grid)
    exec("emo_"+ candidate + "_grid = emo_grid")
    total = np.sum(emo_grid)
    norm_grid = emo_grid/total
    exec("norm_"+ candidate + "_grid = norm_grid")

#create mean grid
total_grid = norm_sanders_grid + norm_clinton_grid
total_grid = total_grid/2.0


#subtract mean grid from individual grids
flat_sanders_grid = norm_sanders_grid-total_grid
flat_clinton_grid = norm_clinton_grid-total_grid


################################################
# Plot 2x1 comparison of differential ratings  #
################################################

fig, axes = plt.subplots(nrows=1, ncols=2, figsize=(10, 4.4))

#standard deviation of variations from 0 (=mean)
sig_tot = np.std([flat_sanders_grid,flat_clinton_grid])

#set font and increase font size
mpl.rcParams.update({'font.size': 17, 'font.family':'sans-serif'})


#first subplot (Clinton)
ax1 = plt.subplot(1, 2, 1)

#show Trump difference map with fixed z-range
im1 = ax1.imshow(flat_clinton_grid, cmap="seismic")

#set contour levels to multiples of standard deviation
levels1 = np.arange(-20.0,20.0,1.0)*sig_tot
CS1 = ax1.contour(flat_clinton_grid, levels1, linewidths=2, colors='white', linestyles='solid')

#axis labels
ax1.set_xlabel('calm', weight='ultralight')
ax1u = ax1.twiny()
ax1u.set_xlabel('excited', weight='ultralight')
ax1.set_xticks([])
ax1u.set_xticks([])
ax1.set_ylabel('unpleasant', weight='ultralight', rotation=0, labelpad=50, ha='center', va='center')
ax1r = ax1.twinx()
ax1r.set_ylabel('')
ax1.set_yticks([])
ax1r.set_yticks([])

#label
ax1.text(5,10, 'Clinton', style='italic', weight='roman')


#second subplot (Sanders)
ax2 = plt.subplot(1, 2, 2)

im2 = ax2.imshow(flat_sanders_grid, cmap="seismic")
levels2 = np.arange(-20.0,20.0)*sig_tot
CS2 = ax2.contour(flat_sanders_grid, levels2, linewidths=2, colors='white', linestyles='solid')

ax2.set_xlabel('calm', weight='ultralight')
ax2u = ax2.twiny()
ax2u.set_xlabel('excited', weight='ultralight')
ax2.set_xticks([])
ax2u.set_xticks([])
ax2.set_ylabel('')
ax2r = ax2.twinx()
ax2r.set_ylabel('pleasant', weight='ultralight', rotation=0, labelpad=42.5, ha='center', va='center')
ax2.set_yticks([])
ax2r.set_yticks([])

ax2.text(5,10, 'Sanders', style='italic', weight='roman')

#save figure in dir 'plots'
plt.savefig('plots/difference_maps_democrats.png', format='png', dpi=150, bbox_inches='tight')