Ahh! It’s Viral!

12 of the Strangest Weather-Related Photographs Ever Taken . Wow . Watch These Rhinos Fly !

Viral and attention-grabbing headlines are certainly intriguing. I definitely fell for clickbait countless times! I suspect there is a formula to generating eye-catching, click-worthy headlines. Wouldn't it be great if we leveraged algorithms to tease apart some of this hidden structure within viral headlines? Well, this is what this exploration is all about. :)

The goal of the notebook is to demonstrate creation of simple Markov chain to generate viral headlines using real examples to build its vocabulary.

Firstly, a big thank you to the wonderful folks over at Ripenn for providing a neat corpus. Data is available for download at end of this post: "7 Things Marketers Can Learn From 2,616 Viral Headlines". I discovered this dataset which was referenced in Buffersocial's post "How to Write The Perfect Headline: The Top Words Used in Viral Headlines" .

My inspiration for building a Markov chain originated from these two posts:

• "Generating pseudo random text with Markov chains using Python"
• "How to fake a sophisticated knowledge of wine with Markov Chains"

Thanks to Shabda Raaj and Tony Fischetti. I learned a great amount from these demonstrations, they're definitely worth a visit. In my ipython notebook, I assimilated some of their techniques and with my own experience working with text data.

notebook (viral-markov-generator.ipynb)