Buckets full of Books

Social media phenomenon are like religious rituals, you might not willingly partake in them but you might find those who you share your life and planet with, involve you nonetheless. The pay it forward like model of the ALS Ice Bucket Challenge is something that those with digital personas are still reeling from. Watching countless videos of celebrities pour water over themselves might fan the perversions of some while its value as a discussion topic is not inconsiderable. ALS and its awareness remained on the back burner before the risk of appearing buffoon like due to the ignorance of the disease despite having doused yourself in freezing H2O became apparent. As big as a problem ALS might be, it is certainly not the biggest in the world. While nay Sayers are hushed by the populous with the rationale that it is better than sitting home and watching Kapil Sharma on YouTube, it does have the potential of good, I still wonder. While the idea remains donate or drench, it is the drenching that draws in likes, comments and social admiration. In India especially it seems ironic that while side stepping beggars, finding eunuchs repulsive and refusing to increase the pay of your housemaid so that her daughter can stay in school, the average possessor of a smart  phone finds comfort in an act of selfless, Cross continental giving. 

However one can’t deny the PR success of the campaign. Why things go viral and others don’t is an interesting line to pursue with a case study on the BucketChallenege for the Ice bucket spawned other non committal, pseudo intellectual challenges like the Books Challenge.  A raging phenomenon this asks you to list out the 10 most influential/inspiring books you have read in your life time, then of course challenge a bunch of friends to do the same. This is more Facebook centric given the demand for a lengthy post. As this article on First Post points out the entire exercise has led to interesting results. Our concern with public perception is natural and requires the propagation of an inflated image to allow for the natural reduction by societal skepticism. That is if your perceived intelligence be measurable in SI units, if it lies around 5 on a scale of 10, you project an 8, your friends are likely to perceive you as a 6.5-7 . It is thus that this thrown down gauntlet is being considered more worthy than doling out money for a disease whose abbreviation’s expansion still remains alien. Philosophy, decidedly obscure and foreign authors, with Children’s books thrown in for authenticity and the creative indulgence into fantasy and comical books, that’s what’s filling my Facebook Wall.

With the intent shifting from including truly inspirational books to making truly sharable and trendable lists, the trend has run the inevitable course of generating annoyance by excessive occurrence. With the proud proclamation of Harry Potter books and the LOTR trilogy (There are 7 books) the unveiled desire for coolness raises its ugly head. One can only be thankful for the masses having refrained from being truly true to their first ever novels and not mentioning that radioactive producer of Literary Litter, Mr Bhagat.   Perhaps there is more to the list of literary works, more to be interpreted from it?

Facebook data analysis

Although the phenomenon is widespread and seemingly trivial, can it be tapped for some insights? If yes then how and what. As we are limited to Facebook with this idea, the Graph Search API comes to mind. With FQL deprecated, Graph Search via the Facebook Graph API explorer offers a neat way of exploring your personal FB Graph and also fetching some basic public data. You could directly query for the count of hashtags using the Hashtags Counter API. This needs APP access so you’d need to register an APP with an AuthCode etc. Open the Explorer, choose APP authentication and use something like:

hashtag_counts?hashtags[]=booksbucketchallenge&since=1400716800

This should get you an idea of how popular the phenomenon is for a particular period allowing for plots of increasing popularity by mention alone over time. This is across visibility constraints set by FB.

If the data suggests that mining more could be beneficial, i.e a lot of people took the challenge, then go for textual search. Post based searches are deprecated for the new Versions of the FB Api. However you can still query the old versions (< V 2.0) for 90-95% of the public results. The qurey shall look something like this (basic user permissions should suffice).

search?q=%23booksbucketchallenge

What this returns is a JSON with a message and message tag components. The pages of the Author’s public profile or the book’s might get attached and you could choose to ignore this data for all you need are the names.

"message": "#BooksBucketChallenge I was nominated by Sunny Gupta for this challenge\n\nHere are 10 books that I believe have heavy influence in my life\n1. The art of war -Sun Tzu\n2. How to win and influence people - Dale Carnegie\n3. The intelligent Investor -Benjamin Graham \n4. Damodaran on Valuation -Professor Aswath damodaran\n5. Rich Dad Poor Dad -Robert Kiyosaki\n6. Life of Pi -Yann Martel\n7. How the Economic Machine Works -Ray Dalio\n8. The Alchemist - Paulo Coehlo\n9. How to lie with statistics -Darell Huff\n10. The Monk who sold his ferrari -Robin Sharma\n\nI nominate Neera Jain, Ankit Shukla and Ankit Kumar for this challenge!"

Extract from the above patterns like <value>-<value> and you get your 10 books. How is this helpful ? Well you could compile a DB with these entries, adding counts for each book. These shall be mostly public posts so you needn’t require access to the user’s likes to know what he or she liked reading. Mapping this against ISBN dbs shall eliminate indexing of misspellings and duplicates. So what you’d end up is with the kind of books people around you or in public claimed to have found very interesting. You avoid the pitfalls of sentiment analysis as these posts by nature are an expression of extreme fondness of said books. Imagine the possibilities of tracking various variables like :

genres, authors, publishers, length, language, date of publishing, country of Author’s origin etc for the book cross referenced with the reader’s : age, gender, country, profession, orientation, other likes. 

Surely such information can be valuable for publishers, authors, avid readers, librarians, students of literature and its impact on society.