The Columbia Business Analytics Blog is a forum for discussing current topics in Business Analytics.

Most Recent Posts



Sites of Interest

  • Pothole Analytics

    Potholes.  Like rude cab drivers, crowded subway cars, and “no right turn on red,” they seem like just another annoying aspect of getting around New York City.  And Manhattan potholes are legendary – as Dave Letterman quipped, “There is a pothole so big on Eighth Avenue that it has its own Starbucks inside.”   The City of New York repaired more 241,500 potholes in 2013 – a truly amazing number.   And the position of the city regarding potholes is that they are an inevitable side-effect of harsh winter weather and that nothing can be done about them other than repairing them after they occur.  But an analysis by Lucius Riccio of the Columbia Business School published in OR/MS Today suggests that this is not true – potholes can be prevented and it is probably economic to do so. 

  • Where the Analytics Jobs Are

    Everyone reading this blog knows that “business analytics” is a hot hiring area right now.  And most people probably have a hunch that a lot of these jobs are in the San Francisco Bay Area.  Most people (at least me) would guess that other “tech-focused cities” such as Seattle and Austin would have more than their share of tech jobs.  In a recent issue of Analytics magazine, Army operations research analyst Scott Nestler published an analysis of 11,584 job listings on LinkedIn on Feb 1, 2014 that contained the keyword “analytics”.  He then mapped the locations of the jobs into the 574 Primary Statistical Areas (PSAs) in the United States.  The top 20 markets for analytics jobs are shown below:

  • k-Nearest Neighbors and The New York Times Most Popular Piece of Content in 2013

    In the Business Analytics class at the Columbia Business School, we teach the k-Nearest Neighbors algorithm in the context of recommendation systems such as those provide by Netflix and Pandora.  But an interesting recent post at Northwestern University’s Knight Lab showed how it was used to help create the dialect heat maps (such as the one shown above) that were the most popular piece of content at the New York Times last year.

  • Bloomberg, Twitter, and the Social Network Data Value Chain

    For almost a year, Bloomberg has offered a service that allows its subscribers to see when mentions of a company spike on Twitter.  Last month, Bloomberg announced a sentiment analysis tool that analyzes the contents of the tweets to determine whether the news is likely to be good or bad for the stock price.   The graphic above shows a screenshot after a CNBC reporter tweeted the news that Comcast would be acquiring Time Warner.    The purple bars are numbers of tweets – the green bars are positive mentions, the orange negative.   

  • Sort-of “Big Data” and College Admissions

    A recent NPR Marketplace posting announced (or at least implied) that the “era of Big Data” had arrived for the college application process:

    ’As a student goes through the search and application process, many times unbeknownst to them, colleges are collecting information about everything that they do,’ says David Hawkins, director of public policy and research for the National Association for College Admission Counseling.  Messages, campus visits, even social media interactions, are logged into admissions software.

    At some level, this is no different from what consumer products and services companies have been doing for some time  -- using “big data” and business analytics to determine which customers to target.  However, the college admissions game has a special twist that sets it apart:

  • Yet Another Online Music Entrant

    Am I the only person around who is not planning to start an online music company this year?  According to an article in the New York Times, Lior Cohen (above) , “one of music’s most respected power brokers”, is planning to start – what else – an online music company called 300.  The twist?  300 will partner with Twitter to help use Twitter’s vast trove of information to help identify new artists and trends.  According to the Times, music is the most popular topic on Twitter – users sent over one billion music-related messages last year.

  • Beats Enters the Online Music Fray

    Another week, another entry into streaming music.   Beats Music launched its online music streaming service that works with AT&T.  While there have been a substantial number of new streaming services popping up over the past few months, as this Time magazine article explains, Beats is more likely than most to make a success of it.  For one thing, the company is massively well funded.  For another thing, the company was founded by Jimmy Lovine and Dr. Dre, who have a long track record of successful music-based collaborations.  Finally – and perhaps most importantly -- Beats has entered into an an agreement with AT&T to serve as a distributor of their service.  Starting January 21, Beats has an introductory offer by which AT&T subscribers can use their service for free.

    So – how does Beats determine what music to recommend?  In a blog post, Beats Music CEO Ian Rogers pokes fun at algorithmic recommendation systems:

  • Fly Before You Buy?

    You’ve heard of “try before you buy” -- now Amazon is considering “fly before you buy”. 

  • Bringing Business Analytics into the Human Resources Department

    Historically, the major consumers of business analytics within most corporations have been finance, marketing, and operations.  However, as two recent articles describe, analytics is increasingly being applied within human resources departments.  These applications even have a name:  workforce analytics.  As a recent article in The Atlantic explains:

    “You can now find dedicated analytics teams in the human-resources departments of not only huge corporations such as Google, HP, Intel, General Motors, and Proctor & Gamble, to name just a few, but also companies like McKee Foods, the Tennessee-based maker of Little Debbie snack cakes.”

  • Decoding Netflix's Genome

    In a recent Atlantic Monthly article, Alex Madrigal describes how he was able to “decode Netflix’s movie genome”.  A recent blog post described Pandora’s music genome project and how it is used to classify every song in their data base according to a list of 250 attributes.  Netflix uses a similar process to classify movies into a large number of categories or subgenres


    “Using large teams of people specially trained to watch movies, Netflix deconstructed Hollywood. They paid people to watch films and tag them with all kinds of metadata. This process is so sophisticated and precise that taggers receive a 36-page training document that teaches them how to rate movies on their sexually suggestive content, goriness, romance levels, and even narrative elements like plot conclusiveness.