11: Visualizations

Distribution of Publications by Author: This bar graph indicates a relatively even distribution of article publications among various authors, with some slight variations. Authors like Field Level Media and The Associated Press appear to be the most prolific, suggesting their dominant role in content creation within the dataset. This can imply a significant influence of these entities on the media landscape covered in the study.

import pandas as pd
data=pd.read_csv('main.csv')
# Checking for missing values and summarizing the data by authors
import plotly.express as px

missing_values = data.isnull().sum()
author_distribution = data['AUTHOR'].value_counts()

author_distribution
fig = px.bar(author_distribution, 
             x=author_distribution.index, 
             y=author_distribution.values, 
             labels={'x':'Author', 'y':'Count'},
             title='Distribution of Publications by Author')
fig.show()

Analyzing Temporal Trends in Article Publications: The line graph shows the number of articles published over time, highlighting noticeable spikes and troughs. These peaks may correspond to major news events or seasonal coverage spikes, suggesting a reactive nature of publications to current events. The data visualization effectively captures the dynamic nature of news cycles, offering insights into how external events drive media output ## Analyzing Temporal Trends in Article Publications

This code converts the ‘DATE’ column to datetime format and groups the data by date to count the number of articles. It then plots these counts over time using a line graph to reveal trends.

import plotly.express as px
# Convert the 'DATE' column to datetime format
data['DATE'] = pd.to_datetime(data['DATE'])

# Group by Date and count the number of articles
date_counts = data.groupby('DATE').size().reset_index(name='counts')

# Plotting using Plotly
fig = px.line(date_counts, x='DATE', y='counts', title='Temporal Trends of Articles', labels={'counts': 'Number of Articles'})
fig.update_layout(xaxis_title='Date', yaxis_title='Number of Articles')
fig.show()

Monthly Publication Trends by Author

This code segments the dataset by month and author, creating a stacked bar chart to visualize the number of articles published by each author over time. It effectively illustrates trends and contributions by individual authors.

# Group by both Date and Author
import pandas as pd
import plotly.graph_objects as go

# Load the dataset and prepare it
data = pd.read_csv('main.csv')
data['DATE'] = pd.to_datetime(data['DATE'])

# Resample data to monthly frequency
data['MONTH'] = data['DATE'].dt.to_period('M')

# Group by both Month and Author
author_month_counts = data.groupby(['MONTH', 'AUTHOR']).size().unstack(fill_value=0)

# Create the figure using Plotly
fig = go.Figure()
for author in author_month_counts.columns:
    fig.add_trace(go.Bar(x=author_month_counts.index.astype(str), y=author_month_counts[author], name=author))

fig.update_layout(
    barmode='stack',
    title='Number of Articles by Author (Monthly)',
    xaxis_title='Month',
    yaxis_title='Number of Articles',
    xaxis={'type': 'category'},  # Change this to 'category' to handle the period index as categorical
    legend_title='Author',
    legend=dict(orientation="h", yanchor="bottom", y=1.02, xanchor="right", x=1)
)
fig.show()

N-grams Analysis (Unigrams, Bigrams, Trigrams, Quadgrams) Unigrams: Most frequently occurring words like “trump,” “new,” and “people” suggest a focus on political figures and general news content. This might indicate prevalent themes in the data or common topics of interest.

Bigrams and Trigrams: Sequences such as “Donald Trump,” “White House,” and “episode people Wednesday” show common phrases that likely relate to specific events or recurring topics, useful for understanding contextual trends.

Quadgrams: More complex phrases like “president Donald Trump said” and “secretary state Rex Tillerson” point towards specific news narratives, helping identify detailed storylines or recurring news frameworks.

import pandas as pd
from sklearn.feature_extraction.text import CountVectorizer
import plotly.express as px
import string

# Load the dataset
data = pd.read_csv('main.csv')

# Function to clean and prepare text
def clean_text(text):
    # Convert to lower case and remove punctuation
    text = text.lower()
    text = text.translate(str.maketrans('', '', string.punctuation))
    return text

# Apply text cleaning
data['cleaned'] = data['cleaned'].apply(clean_text)

# Function to create n-gram frequency plot
def plot_ngrams(n, max_features=26):
    # Create CountVectorizer object for n-grams
    vectorizer = CountVectorizer(ngram_range=(n, n), stop_words='english', max_features=max_features)
    
    # Fit and transform the data
    ngrams = vectorizer.fit_transform(data['cleaned'])
    
    # Sum up their counts and get feature names
    sum_ngrams = ngrams.sum(axis=0) 
    ngrams_freq = [(word, sum_ngrams[0, idx]) for word, idx in vectorizer.vocabulary_.items()]
    
    # Sort n-grams by frequency
    ngrams_freq = sorted(ngrams_freq, key=lambda x: x[1], reverse=True)
    words, freqs = zip(*ngrams_freq)
    
    # Create a DataFrame
    df_ngrams = pd.DataFrame({'N-gram': words[1:], 'Frequency': freqs[1:]})
    
    # Plot using Plotly
    if n==1:
        grams='Uni'
    elif n==2:
        grams='Bi'
    elif n==3:
        grams='Tri'
    elif n==4:
        grams='Quad'
    fig = px.bar(df_ngrams, x='N-gram', y='Frequency', title=f'Top {25} {grams}-grams', 
                 template='plotly_dark', color='Frequency', color_continuous_scale=px.colors.sequential.Viridis)
    fig.show()

# Plot unigrams, bigrams, and trigrams

Unigrams

plot_ngrams(1)

Bigrams

plot_ngrams(2)

Trigrams

plot_ngrams(3)

Quadgrams

plot_ngrams(4)