File size: 11,836 Bytes
21ac434
7bf8be4
 
 
 
 
 
 
 
 
 
 
20f957a
93b4f33
 
 
 
 
 
20f957a
68a2713
72c555f
 
21b510b
72c555f
 
21b510b
72c555f
 
 
 
68a2713
93b4f33
 
 
7bf8be4
b7e92a0
 
928cd59
b7e92a0
 
21b510b
 
 
 
 
68a2713
 
 
 
 
 
20f957a
68a2713
 
 
 
7bf8be4
68a2713
 
7bf8be4
 
20f957a
 
 
928cd59
7bf8be4
 
 
20f957a
 
 
 
928cd59
20f957a
 
 
7bf8be4
20f957a
 
 
928cd59
7bf8be4
 
20f957a
 
848f63d
 
 
 
 
 
 
 
 
 
 
 
928cd59
848f63d
 
 
7bf8be4
836b08d
 
 
 
 
 
 
 
 
 
928cd59
836b08d
 
 
 
 
 
 
 
 
 
928cd59
836b08d
 
928cd59
 
 
 
 
 
 
 
 
 
 
836b08d
 
 
 
 
928cd59
836b08d
 
 
 
 
 
 
 
 
 
928cd59
836b08d
 
7bf8be4
72c555f
 
 
928cd59
7bf8be4
 
 
 
 
 
928cd59
7bf8be4
 
21b510b
20f957a
928cd59
 
5ecc9ee
7bf8be4
20f957a
7bf8be4
20f957a
 
5ecc9ee
20f957a
 
 
848f63d
 
 
 
 
20f957a
 
 
5ecc9ee
836b08d
 
 
 
20f957a
5ecc9ee
836b08d
 
928cd59
 
90742aa
20f957a
 
5ecc9ee
836b08d
 
 
90742aa
20f957a
 
 
 
 
72c555f
20f957a
 
7bf8be4
848f63d
21b510b
d44cfa6
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
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
import streamlit as st
import pandas as pd
import matplotlib.pyplot as plt
from sklearn.feature_extraction.text import CountVectorizer
import seaborn as sns
import plotly.express as px
import plotly.io as pio
import plotly.graph_objects as go

# Set page configuration
st.set_page_config(layout="wide")

# Function to load and clean data
def load_and_clean_data():
    df1 = pd.read_csv("data/reviewed_social_media_english.csv")
    df2 = pd.read_csv("data/reviewed_news_english.csv")
    df3 = pd.read_csv("data/tamil_social_media.csv")  
    df4 = pd.read_csv("data/tamil_news.csv")       

    # Concatenate dataframes and clean data
    df_combined = pd.concat([df1, df2, df3, df4])
    
    # Replace 'nan' and 'None' with numpy NaN for removal
    df_combined['Domain'] = df_combined['Domain'].replace({"MUSLIM": "Muslim", "nan": pd.NA, "None": pd.NA, "Other-Ethnic": "Other-Ethnicity"})
    
    # Specific replacements for 'Sentiment' and 'Discrimination'
    df_combined['Sentiment'] = df_combined['Sentiment'].replace({"nan": pd.NA, "None": pd.NA, "No": pd.NA})
    df_combined['Discrimination'] = df_combined['Discrimination'].replace({"nan": pd.NA, "None": pd.NA, "No": pd.NA})
    
    # Drop rows with NA values in 'Domain', 'Sentiment', and 'Discrimination'
    df_combined.dropna(subset=['Domain', 'Sentiment', 'Discrimination'], inplace=True)

    return df_combined

df = load_and_clean_data()


# Page navigation setup
page_names = [" GESI Overview", "Sentiment Analysis", "Discrimination Analysis", "Channel Analysis"]
page = st.sidebar.selectbox("Choose a page", page_names)

# Sidebar Filters
domain_options = df['Domain'].dropna().unique()
channel_options = df['Channel'].dropna().unique()
sentiment_options = df['Sentiment'].dropna().unique()
discrimination_options = df['Discrimination'].dropna().unique()

domain_filter = st.sidebar.multiselect('Select Domain', options=domain_options, default=domain_options)
channel_filter = st.sidebar.multiselect('Select Channel', options=channel_options, default=channel_options)
sentiment_filter = st.sidebar.multiselect('Select Sentiment', options=sentiment_options, default=sentiment_options)
discrimination_filter = st.sidebar.multiselect('Select Discrimination', options=discrimination_options, default=discrimination_options)

# Apply filters
df_filtered = df[(df['Domain'].isin(domain_filter)) & 
                 (df['Channel'].isin(channel_filter)) & 
                 (df['Sentiment'].isin(sentiment_filter)) & 
                 (df['Discrimination'].isin(discrimination_filter))]

# Define a color palette for consistent visualization styles
color_palette = px.colors.sequential.Viridis


# Visualisation for Domain Distribution
def create_pie_chart(df, column, title):
    fig = px.pie(df, names=column, title=title, hole=0.35)
    fig.update_layout(margin=dict(l=6, r=6, t=12, b=6), legend=dict(x=0.1, y=1), font=dict(size=7)
    fig.update_traces(marker=dict(colors=color_palette))
    return fig

# Visualization for Distribution of Gender versus Ethnicity
def create_gender_ethnicity_distribution_chart(df):
    df['GenderOrEthnicity'] = df['Domain'].apply(lambda x: "Gender: Women & LGBTQIA+" if x in ["Women", "LGBTQIA+"] else "Ethnicity")
    fig = px.pie(df, names='GenderOrEthnicity', title='Distribution of Gender versus Ethnicity', hole=0.35)
    fig.update_layout(margin=dict(l=6, r=6, t=12, b=6), legend=dict(x=0.1, y=1), font=dict(size=7)
    return fig

# Visualization for Sentiment Distribution Across Domains
def create_sentiment_distribution_chart(df):
    df['Discrimination'] = df['Discrimination'].replace({"Non Discriminative": "Non-Discriminative"})  # Assuming typo in the original script
    domain_counts = df.groupby(['Domain', 'Sentiment']).size().reset_index(name='counts')
    fig = px.bar(domain_counts, x='Domain', y='counts', color='Sentiment', title="Sentiment Distribution Across Domains", barmode='stack')
    fig.update_layout(margin=dict(l=6, r=6, t=12, b=6), xaxis_title="Domain", yaxis_title="Counts", font=dict(size=7))
    return fig

# Visualization for Correlation between Sentiment and Discrimination
def create_sentiment_discrimination_grouped_chart(df):
    # Creating a crosstab of 'Sentiment' and 'Discrimination'
    crosstab_df = pd.crosstab(df['Sentiment'], df['Discrimination'])
    
    # Check if 'Yes' and 'No' are in the columns after the crosstab operation
    value_vars = crosstab_df.columns.intersection(['Yes', 'No']).tolist()
    
    # If 'No' is not in columns, it will not be included in melting
    melted_df = pd.melt(crosstab_df.reset_index(), id_vars='Sentiment', value_vars=value_vars, var_name='Discrimination', value_name='Count')
    
    # Proceeding to plot only if we have data to plot
    if not melted_df.empty:
        fig = px.bar(melted_df, x='Sentiment', y='Count', color='Discrimination', barmode='group', title="Sentiment vs. Discrimination")
        fig.update_layout(margin=dict(l=6, r=6, t=12, b=6), xaxis_title="Sentiment", yaxis_title="Count", font=dict(size=7))
        return fig
    else:
        return "No data to display for the selected filters."

# Function for Top Domains with Negative Sentiment Chart
def create_top_negative_sentiment_domains_chart(df):
    domain_counts = df.groupby(['Domain', 'Sentiment']).size().unstack(fill_value=0)
    domain_counts.sort_values(by='Negative', ascending=False, inplace=True)
    domain_counts_subset = domain_counts.iloc[:3, [0]]
    domain_counts_subset = domain_counts_subset.rename(columns={domain_counts_subset.columns[0]: 'Count'})
    domain_counts_subset = domain_counts_subset.reset_index()
    colors = ['limegreen', 'crimson', 'darkcyan']
    fig = px.bar(domain_counts_subset, x='Count', y='Domain', title='Top Domains with Negative Sentiment', color='Domain',
                 orientation='h', color_discrete_sequence=colors)
    fig.update_layout(margin=dict(l=6, r=6, t=12, b=6), xaxis_title="Negative sentiment content Count", yaxis_title="Domain", font=dict(size=7))
    return fig

# Function for Key Phrases in Negative Sentiment Content Chart
def create_key_phrases_negative_sentiment_chart(df):
    cv = CountVectorizer(ngram_range=(3,3), stop_words='english')
    trigrams = cv.fit_transform(df['Content'][df['Sentiment'] == 'Negative'])
    count_values = trigrams.toarray().sum(axis=0)
    ngram_freq = pd.DataFrame(sorted([(count_values[i], k) for k, i in cv.vocabulary_.items()], reverse=True))
    ngram_freq.columns = ['frequency', 'ngram']
    fig = px.bar(ngram_freq.head(10), x='frequency', y='ngram', orientation='h', title='Key phrases in Negative Sentiment Content')
    fig.update_layout(margin=dict(l=6, r=6, t=12, b=6), xaxis_title="Frequency", yaxis_title="Trigram", font=dict(size=7))
    return fig

# Function for Key Phrases in Positive Sentiment Content Chart
def create_key_phrases_positive_sentiment_chart(df):
    cv = CountVectorizer(ngram_range=(3, 3), stop_words='english')
    trigrams = cv.fit_transform(df['Content'][df['Sentiment'] == 'Positive'])
    count_values = trigrams.toarray().sum(axis=0)
    ngram_freq = pd.DataFrame(sorted([(count_values[i], k) for k, i in cv.vocabulary_.items()], reverse=True))
    ngram_freq.columns = ['frequency', 'ngram']
    fig = px.bar(ngram_freq.head(10), x='frequency', y='ngram', orientation='h', title='Key phrases in Positive Sentiment Content')
    fig.update_layout(margin=dict(l=6, r=6, t=12, b=6), xaxis_title="Frequency", yaxis_title="Trigram", font=dict(size=7))
    return fig
    
# Function for Prevalence of Discriminatory Content Chart
def create_prevalence_discriminatory_content_chart(df):
    domain_counts = df.groupby(['Domain', 'Discrimination']).size().unstack(fill_value=0)
    fig = px.bar(domain_counts, x=domain_counts.index, y=['Discriminative', 'Non-Discriminative'], barmode='group',
                 title='Prevalence of Discriminatory Content')
    fig.update_layout(margin=dict(l=6, r=6, t=12, b=6), xaxis_title="Domain", yaxis_title="Count", font=dict(size=7))
    return fig

# Function for Top Domains with Discriminatory Content Chart
def create_top_discriminatory_domains_chart(df):
    domain_counts = df.groupby(['Domain', 'Discrimination']).size().unstack(fill_value=0)
    domain_counts.sort_values(by='Discriminative', ascending=False, inplace=True)
    domain_counts_subset = domain_counts.iloc[:3]
    domain_counts_subset = domain_counts_subset.rename(columns={'Discriminative': 'Count'})
    fig = px.bar(domain_counts_subset, x='Count', y=domain_counts_subset.index, orientation='h',
                 title='Top Domains with Discriminatory Content')
    fig.update_layout(margin=dict((l=6, r=6, t=12, b=6), xaxis_title="Discriminatory Content Count", yaxis_title="Domain", font=dict(size=7))
    return fig

# Function for Channel-wise Sentiment Over Time Chart
def create_sentiment_distribution_by_channel_chart(df):
    sentiment_by_channel = df.groupby(['Channel', 'Sentiment']).size().reset_index(name='counts')
    fig = px.bar(sentiment_by_channel, x='Channel', y='counts', color='Sentiment', title="Sentiment Distribution by Channel", barmode='group')
    fig.update_layout(margin=dict(l=6, r=6, t=12, b=6), xaxis_title="Channel", yaxis_title="Counts", font=dict(size=7))
    return fig

# Function for Channel-wise Distribution of Discriminative Content Chart
def create_channel_discrimination_chart(df):
    channel_discrimination = df.groupby(['Channel', 'Discrimination']).size().unstack(fill_value=0)
    fig = px.bar(channel_discrimination, x=channel_discrimination.index, y=['Discriminative', 'Non-Discriminative'], barmode='group')
    fig.update_layout(title='Channel-wise Distribution of Discriminative Content', margin=dict(l=6, r=6, t=12, b=6), font=dict(size=7))
    return fig

# Function for rendering dashboard
def render_dashboard(page, df_filtered):
    if page == " GESI Overview":
        st.title(" GESI Overview Dashboard")
        col1, col2 = st.columns(2)
        with col1:
            st.plotly_chart(create_pie_chart(df_filtered, 'Domain', 'Distribution of Domains'))
        with col2:
            st.plotly_chart(create_gender_ethnicity_distribution_chart(df_filtered))

        col3, col4 = st.columns(2)
        with col3:
            st.plotly_chart(create_sentiment_distribution_chart(df_filtered))
        with col4:
            chart = create_sentiment_discrimination_grouped_chart(df_filtered)
            if isinstance(chart, str):
                st.write(chart)
            else:
                st.plotly_chart(chart)

    elif page == "Sentiment Analysis":
        st.title("Sentiment Analysis Dashboard")
        col1, col2 = st.columns(2)
        with col1:
            st.plotly_chart(create_sentiment_distribution_chart(df_filtered))
        with col2:
            st.plotly_chart(create_top_negative_sentiment_domains_chart(df_filtered))

        col3, col4 = st.columns(2)
        with col3:
            st.plotly_chart(create_key_phrases_negative_sentiment_chart(df_filtered))
        with col4:
            st.plotly_chart(create_key_phrases_positive_sentiment_chart(df_filtered)

    elif page == "Discrimination Analysis":
        st.title("Discrimination Analysis Dashboard")
        col1, col2 = st.columns(2)
        with col1:
            st.plotly_chart(create_prevalence_discriminatory_content_chart(df_filtered))
        with col2:
            st.plotly_chart(create_top_discriminatory_domains_chart(df_filtered))

    elif page == "Channel Analysis":
        st.title("Channel Analysis Dashboard")
        col1, col2 = st.columns(2)
        with col1:
            st.plotly_chart(create_sentiment_distribution_by_channel_chart(df_filtered))
        with col2:
            st.plotly_chart(create_channel_discrimination_chart(df_filtered))


# Render the selected dashboard page
render_dashboard(page, df_filtered)