saturn

This dataset is an index of 2,103 articles from The Onion, a satirical news outlet, containing article headlines, thumbnail image URLs, and a sequential row ID. The headline column is the most analytically interesting field, with a vocabulary of 7,613 unique words and mean headline length of about 9 words and 61 characters — worth exploring for length distribution and common vocabulary patterns. There are also 209 duplicate image URLs (~10% of rows), suggesting some thumbnails are reused across multiple articles, with one image appearing 11 times.

This column contains Kinja/Gawker CDN image URLs, serving as thumbnail or article image references across 2103 rows. Every value is a URL (url_rate 1.0) pointing to a fixed-size JPEG (w_645, q_60 transform parameters are baked into all URLs), making the path prefix structurally identical across all rows with only the hash filename varying. Surprising: 209 duplicate URLs exist (duplicate_rate ~9.9%), with one image URL appearing 11 times, suggesting multiple articles or rows share the same thumbnail — likely a placeholder or default image reuse pattern.

This column contains satirical or parody news headlines, with the column name itself being a representative example ('That'll Be $3,' Says Trump After Handing Water Bottle To Sick Ohio Resident). With 2095 unique values out of 2103 rows and a mean length of ~61 characters (~9 words), these are near-unique short text strings consistent with headline-style content. Surprisingly, 8 duplicate headlines exist despite the near-unique alert, and the Flesch readability score of 46.87 indicates moderately difficult prose — higher than typical clickbait but consistent with crafted satirical writing. Top words ('to', 'of', 'in', 'new') are generic function words offering little discriminative signal on their own.

This column is almost certainly a row index or sequential identifier: it contains 2103 values, all unique, with min=2 and max=2104, perfectly symmetric distribution (skew=0.0, mean=median=1053.0), and zero outliers. The near-perfect uniformity (platykurtic at -1.2, IQR=1051 spanning exactly half the range) confirms an arithmetic sequence. The column name '1' suggests it was emitted as an unnamed index column during export.