How to quickly fill the vacancy of large data talents

We should stop trumpeting the gap between big data technicians and start talking about training and recruiting programs.

Experts are trained, not born. One question you need to answer is whether you need technical expertise or data expertise. Technical expertise allows SQL databases or Hadoop clusters to be put into use. The expertise of data enables us to use data mining, text mining, forecasting, and machine learning techniques. Either way, we should first assume that nurturing talent from existing employees is the most important and the best choice.

The 2012 large data and Analysis personnel survey showed that most companies plan to take a combination of retraining, recruitment and outsourcing to fill the talent gap. But it's a time-consuming process to get approval for a salary hike, work with new recruits, and interview candidates. More importantly, professionals probably have a good ability to deal with your business, your data sources and your current analysis, and most likely, analyze what you want from large data and advanced analysis.

If you lack technical expertise, trust your existing database administrators, database developers, and data Warehouse technicians to learn new skills. For example, moving from a traditional database to a large-scale parallel processing (MPP) database platform is not a huge leap for database administrators (DBAs), and good people will quickly adapt to the challenge.

Suppliers always have the option of providing low-cost training and accreditation. HP, for example, has just announced course modules and certification exams around its mpp-based HP Vertica Analytics platform. Its introductory course will be taught in HP Expertone classrooms around the world for a period of over three days. Another option is virtual online training, which will be completed within four days. "It is training on how to set up software and use it most effectively," said Chris Selland, vice president of marketing at Hewlett-Packard, which has a wide audience, from database administrators (DBAs) to data scientists. ”

Large data is largely based on the variety of its data and its large capacity, which operates mainly in new platforms such as Apache Hadoop and NoSQL databases. John Schroeder, chief executive of MAPR, the Hadoop software, support and training service provider, points out that in just five months many companies have not had the experience of putting Hadoop clusters into production. "Now we're going to have a talent scramble because we have more projects on hand than the staff, but we'll also see a big increase in the number of Hadoop training compared to 2010 and 2011," he said. ”

MAPR and its rival, Cloudera and Hortonworks, offer a number of training courses. Here in MapR, they usually offer MAPR learning videos from two to 3rd, and their online and classroom topics include the basics of cluster management, hive and pig hands-on, mapreduce training, and hbase programming. Cloudera University offers online and classroom courses worldwide, while Hortonworks University offers classroom courses in large and medium-sized cities, recruiting 6 or more participants, located within the company.

The premise of these training programmes is that the trainers, who are competent system administrators and developers, can quickly learn new technologies that are completely different from SQL and relational databases. In more advanced data analysis-oriented courses, such as "Getting Started with data science-building a recommendation system" (provided by Cloudera University), the target audience is "software engineers, data analysts, and statisticians who have the knowledge and level of basic Hadoop."

If your data management team is not involved in advanced data mining, text mining, and predictive analysis topics, you may want to consider them participating in one of these top analysis masters degree programs. The shortest training course can be completed within 10 months, but it is full-time and may not be appropriate for existing employees unless they can take leave. There are also part-time and school courses, the shortest time required about two years of training.

A master's degree in development is primarily for those who want to attend training, and if you are eager to fill a vacancy, you can think of these courses as a resource for recruitment. The advantage of taking a master's course is that most students will have three or three years of business experience. Most schools are happy to welcome new employers to their recruiting programs, especially those that have a steady demand for new employees.

Some schools even work with sponsoring companies, professors and graduate teams to cope with real-world business challenges. This is the case with the Institute of Advanced Analysis at North Carolina State University and the MIT Digital Research Center. So you can also look for research centers or institutes that offer degree programs with analytical or large data to work together.

The bottom line is that we have a lot of options in addition to complaining or hiring high-priced headhunters to help. If you are looking for a high-end talent, the task is undoubtedly very difficult. But before you send a "I can't find talent" hysteria, determine your needs and the size of the training, and then connect with a local university, perhaps your problem is solved.

(Responsible editor: The good of the Legacy)