To post a job ad, Click here.

Saturday, April 19, 2008

Interview with Ajay Ohri, Data Mining consultant, DecisionStats.com


Ajay Ohri has been working in the field of analytics since 2004 , when it was a still nascent emerging Industries in India. He has worked with the top two Indian outsourcers listed on NYSE,and with Citigroup on cross sell analytics where he helped sell an extra 50000 credit cards by cross sell analytics .He was one of the very first independent data mining consultants in India working on analytics products and domestic Indian market analytics .He regularly writes on analytics topics on his web site www.decisionstats.com and is currently working on open source analytical tools like R and newer softwares like WPS.

Which analytical fields are likely to experience growth, and why?

Analytical fields in fraud detection are likely to grow exponentially as the payment markets grow and frauds become more sophisticated. Another field of analytics that I believe is likely to grow could be environmental analytics especially related to carbon footprints and carbon credits.exchange. Increasingly sophisticated tools for measuring risk in order to cope with current failure of risk management tools to assess risks of leveraged collateral's assets, especially stress testing of such financial products.

Which methodologies might become obsolete, which ones are likely to entertain growth?

Traditional models utilizing offline modes of download data, model, validate, apply scores may face challenges from machine learning algorithms that execute and score regression models in PMML in real time. Some techniques like genetic algorithms and other mixed models will take off. There is likely to more profusion of open source tools like R in the space reserved by SAS and SPSS.Data feeds are likely to be more automated with real time validation rather than one off modeling scoring and periodic validation.

Recommendations for students starting an analytical career or choosing a University curriculum

Pick one macro economics course, one basic quant course , and learn softwares like SPSS , SAS, and especially R using the GUI RCmdr. Improve your spreadsheets skills by learning some basic VBA. If possible try and learn some programming in .Net or create your website using Wordpress.

Play with these, read some blogs on them and have fun playing with numbers. Data Mining and Analytics can reveal remarkable insights and you can be the big man on campus if you do this correctly. Plus the shortage of skilled people will give you an edge over others who dont have these skills to get a job.

Your opinion on outsourcing: impact on US and other economies, how to take advantage of it, reverse outsourcing

Well most Indian outsourcing companies are actually owned by American private equity funds or American companies. It is estimated that for 1 dollar outsourced 69 cents flows back to the host country due to reduced costs,It becomes controversial because essentially its a transfer from blue collar American jobs to white collar American investors.Also countries like India become stable allies and become big market for American products like Microsoft, Pepsi, and Levis.

Reverse outsourcing is already happening. Thats because India is a much cheaper place to live. A 50000 dollar salary (which is so small at home) will get you a plush apartment, club membership, chauffeur driven car, good social life in the embassy circle. There is a big demand for American expatriates especially as Indians find selling/marketing is an unique American strength. Indian people are one of the friendliest people to Americans because of English Education and Hindu domination. Companies like TCS are setting up development centers in America to be closer to the talent.

Remember low end jobs will get outsourced, innovation driven high value jobs will stay in US. Many people I know work in a global delivery model thanks to technologies like skype, remote desktop, etc where it doesn't matter which country team members are.

What are the biggest successes of data mining and statistical sciences in the corporate world, or for humanity in general? What are its biggest failures?

Google is the biggest success in data mining.Pharmaceutical sciences are biggest successes.Biggest failures have been in financial services where focus got lost to sell rather than manage risk as well. Its understandable since penalties for predatory lending are much lower than penalties for wrong medicine products. Even though macro economic pain causes is much more.

Comments on data mining and privacy

Data mining and privacy are surprisingly simple but confounding in number of incidents. A simple encryption of sensitive data can be done using public key and private key before storing the data in tables and datasets. Most analysis is done on non privacy data (names and address rarely get analyzed) hence these can be decrypted only at time of marketing only.

I like Google's approach to opt in/opt out privacy, but am still worried but about the electronic trails leave on the Internet. Maybe there is a demand for a product that deletes all electronic traces in all forums etc..:)

Best practices for analytic professionals: what are the most important items?

Best Practise- Be aware of whats going in your field, subscribe to the blog, email lists, spend at most 10 minutes a day polishing your knowledge or cutting down costs, its fun and refreshes you too. I recommend the SAS -L List, SPSS Google groups and R Help list . You can also add custom Google alerts for specific keywords like Business Intelligence. There are many groups in Linkedin that cater to this segment. In addition, I am a member of Google Chart group and blog on www.decisionstats.com

Above all do network with fellow professionals using online off line tools to stay on your toes in this fast changing field.

Which analytic web sites do you use most or like best? Which analytic products do you use most or like best?

I like the SAS -L list as it really helps in talking about trends etc in analytics. In Websites, I like Mashable.

I love using SAS modules , SPSS add ins, and R packages. Lately I love experimenting a base SAS clone called WPS (http://www.teamwpc.co.uk/products/wps/features/workbench) which is available for only 300 pounds and has a great design

How and why did you become an analytic professional? Which factors contributed to your success?

I became an analytics professional by accident as I joined the team of one of the founder- CEO 's of India's analytical industries at GE 's back office, Shrikant Dash in 2004 when analytics was still very nascent in India. He was was tough customer,being VP, CRM for GE Consumer Finance Americas and he really taught me to work hard in modeling, SAS , reporting.

The rest I learnt just be keeping my eyes and ears open for newer macros, programs ,softwares that help me get the job done faster so I could go home to my family earlier. I found that sticking to basic questions helped me in designing basic analytics products much better than following the software's instructions. Networking aggressively with world class professionals like Alan Churchill of www.savian.net, Phil of www.minequest.com also helped me learn a great deal. A quality I learnt was when in doubt , ask around than fiddle and waste time.

No comments: