Business News Dataset from Google News API

By shrijeetverma13 · May 2, 2026

This dataset contains real-time global business news articles collected using a public API. The raw JSON response has been cleaned and structured into a…

The dataset contains 23 articles from 16 different publishers. The New York Times and TechCrunch are tied at the top with 3 articles each, followed by CNN with 2 articles. A bar chart has been generated showing the full distribution of articles across all publishers.

The analysis shows how average snippet length varies across different publishers. A bar chart and data tables have been generated to visualize these differences, making it easy to compare snippet lengths side by side.

A data table has been generated showing the most frequently mentioned keywords across article titles. The table highlights which words appear most often in the titles after filtering out common stopwords like 'the', 'and', 'is', etc.

A bar chart has been generated showing news publishing activity for each hour of the day (0–23). The visualization reveals how article publishing is distributed throughout the day, making it easy to spot peak and off-peak periods at a glance.

The dataset contains 23 articles published across a roughly 26-hour window, spanning from April 20, 2026 at 12:00 PM to April 21, 2026 at 2:36 PM. Two bar charts were generated showing the hourly and daily distribution of article publishing activity.

The analysis produced two visualizations comparing how different publishers use article title lengths. A bar chart shows the average title length (in characters) for each publisher, ranked from longest to shortest, and a box plot reveals the full distribution of title lengths per publisher — showing not just averages but also variation and outliers.

The analysis examined publishing patterns across 23 articles, grouping them into 15-minute time windows. A bar chart has been generated showing article counts per time bucket, giving a clear visual picture of when news activity spiked. On average, articles were published about 72.5 minutes apart, suggesting a relatively spread-out publishing cadence rather than rapid-fire bursts.

The analysis examined 23 news article URLs, extracting the domain and URL path depth for each. Using the IQR method, the outlier bounds were set between 29 and 153 characters. With a median URL length of 94 characters, all 23 URLs fell comfortably within this range — meaning no outliers were detected. A scatter plot was generated showing URL length on the x-axis and publisher on the y-axis, with point size representing path depth.

The analysis calculated the ratio of title length to snippet length for each publisher and produced two visualizations: a box plot showing the distribution of ratios per publisher (ranked by mean ratio), and a grouped bar chart displaying the mean ratio with standard deviation error bars for each publisher. Publishers with higher ratios use more descriptive titles relative to their snippet previews.

The analysis of news article URLs reveals that all 23 articles in the dataset come from exactly one domain, giving it a 100% share of the content. Both a pie chart and a bar chart have been generated to visualize this distribution, and a data table is also available. The average number of articles per domain is 23.00, since there is only one domain present.

The sentiment analysis of the 'snippet' column reveals an overall positive tone across the dataset. Nearly half of all entries (47.8%) carry positive sentiment, while about 39% are neutral and only 13% are negative. A bar chart has been generated showing the full sentiment distribution visually.

A word cloud has been generated from the 'snippet' column of your dataset. The visualization displays the most frequently occurring words in the text data, with larger words appearing more prominently to indicate higher frequency.