There are several inspirational people in the data analytics arena who with their experience and vision continuously strive to nurture and grow the industry. Dr. Vincent Granville is one such visionary data scientist who has brought much to the field of analytics.
Vincent is a former post-doctorate of Cambridge University and the National Institute of Statistical Sciences. He has had over 15 years of corporate experience, in varied industries including internet, online advertising and finance. His forte though is predictive modeling, digital and business analytics and he is considered the leading expert in scoring technology, fraud detection and web traffic optimization and growth. Vincent has published 40 papers in statistical journals, owns multiple patents, published the first data science book, and raised $6MM in start-up funding. Last year he was listed as one of the top 20 most influential big data experts by Forbes.
In 2008 Vincent founded and self-funded Analyticbridge, a leading community for analytic professionals. More recently, he launched Data Science Central, the leading social network for big data, business analytics and data science practitioners.
In an Analytic bridge interview, reproduced below ( part of a blog post series about data scientists) , writer and blogger Amy gets insights from Vincent about his job, career path, what techniques he uses, corporate culture as well as some advice for data scientists. His journey is indeed inspiring.
Can you summarize your career path in a few sentences?
PhD in computational statistics / image analysis, 1993. Post doctorate (Cambridge UK and NISS, North Carolina), then 15 years of corporate experience ranging from statistician, market research, fraud detection expert, business intelligence, analytics consulting, to Chief Scientist roles. Industries: Internet, online advertising, finance.
Founded and self-funded Analyticbridge, a leading community for analytic professionals, in 2008. Raised 6MM in VC funding in 2006, worked with Wells Fargo, InfoSpace, eBay, Microsoft and a few startups. Own patents on Internet traffic quality scoring (2008-2010), author of the first ebook on data science (2012), listed as top 20 most influential big data experts by Forbes (2012).
How did you find your current job? What skills does it require?
Working exclusively for one employer is dangerous from a risk management point-of-view: it’s like having a stock portfolio with just one stock. Being independent is much more fun, more financially rewarding, and much less risky if done well. Most importantly, I wanted to help people connect and leverage skills by bridging together several communities that are heavy analytic users: statistics, operations research, BI, quant, data mining, six sigma, econometrics etc. The opportunity appeared in 2008, with the emergence and growth of social networks.
My job is similar to being COO / CEO, but I still perform statistical analyses. It requires good jugment, acting quickly and using the right business metrics (to assess long term potential, and success), change management, forecasting, management consulting type of skills, opportunity detection, being aware of new technologies and competitors, taking calculated risks, assessing vendors and partnerships, sales and communications with clients / vendors / users, accounting, a bit of law / finance, and computational marketing (we have a secret recipe for growth). These skills and craftmanships are more important than technical skills learned at school.
What do you enjoy most and least in your current occupation?
I am happy when great partnerships are created thanks to my company, when applicants find jobs thanks to us, when we make clients happy or valuable members successful (some got PhD scholoraships in US thanks to us). I don’t like mundane activities (about 40% of my workload right now), but we are recently doing a better job at outsourcing most of them (accounting, etc.)
What kind of models and data do you use?
I use data gathered on the Internet with web crawlers, or internal unstructured data combined with text mining to produce reports about the analytic community, or to generate stock price forecasts (buy and sell signals for major indices) based on internal job ads (sales) data and other metrics, or testing new patentable algorithms that we describe in our open-source, free data science ebook sponsored by advertisers.
How can someone become a data scientist?
We start to see a few interesting curricula: Northwestern University, Berkeley, Stanford — check our course section on Analyticbridge. Attending these online classes is a good start, as well as being intern for an analytic company (including ourselves). A lot of the technical stuff and resources can be found and learned online or in in our e-book, for free: open-source languages (Python, Hadoop environment, SQL, Java, C++), cross-validation, analytics as a service, model fitting, machine learning, clustering techniques, hidden decision trees, lift metrics, design of experiments, Monte-Carlo simulation, non parametric confidence intervals, association rules and scoring technology etc. We plan to offer a course, and we already offer a free data science certification based on your bio (no exam required).
Another idea is to play with Google keyword API’s and start creating your search engine or your own taxonomies. Attend conferences (check out our conference section). Or download free trial copy of vendor software (Lavastorm, Data Miner, etc.) And buy books on Amazon: read Read our book and journal section on Analyticbridge for details.
What is the corporate culture in your company?
Optimizing all processes by outsourcing as much as possible to carefully selected vendors: mailing list and server management, advertising delivery, social network platform, credit card processing, etc. Making partners, clients and valuable members happy. Killing spam. Offering salaries or hourly rates above average thanks to smart business optimization and the leverage of a number of business “unfair advantages”.