Five great myths about big Data

At present, a large number of enterprises have a huge amount of customer information, including online transactions and social media data. However, the key to success is to be able to gain insights from the data from different sources and channels, while companies with the ability to collect and analyze these data will have a significant advantage in the competition.

However, the unstructured data has become a major challenge for the enterprise. Companies are already familiar with collecting and analyzing structured data, such as traditional sales report information. At present, many enterprises are puzzled how to collect and analyze more types of multiple structured data. such as blog, Radio frequency identification (RFID), sensor networks, social networks, Internet text and documents, Internet search index, detailed call records, medical records, photographic files, video files and E-commerce transaction data.

Due to the structural problems of these data and the complex association of large data types, it is impossible to apply the existing traditional techniques for large data analysis. This brings new tasks to the enterprise and requires developing a new set of methods that can handle not only traditional data, but also easy analysis and application of these emerging data, not just storage.

Myth One: Big data is about the amount of data and data growth

This is not entirely true. Indeed, big data includes massive exponential growth in traditional business data, as well as data generated by new channels such as Web applications, sensor networks, social networks, genomes, videos, and photos. At the same time, large data is still very complex, it is very difficult to collect, store, manage and analyze.

Currently, both types of data are growing. "Companies are drowning in the ocean of information but still eager to get more information, which is a huge opportunity for big data analysis and management," according to the 2011 Top ten forecasts published by IDC Group. The report points out that the company's aspirations will eventually come true. "Global data Volumes (Digital universe) will expand by nearly 50% to about 1.8 gigabytes (about 2 trillion GB). As a reference, experts estimate that 1-gigabyte bytes are equivalent to the amount of data produced by High-definition video files of up to 36 million years in length. ”

Myth Two: Companies should eliminate and replace existing analysis systems to cope with the arrival of the big data age

Wrong, no need! Building large data analysis capabilities requires a perfect combination of talent, process and technology. If the enterprise has not yet discovered the value of the existing business intelligence environment, it is necessary to take the lead in solving the problem before enabling the large data analysis platform. When the traditional business data analysis is endowed with large data, the real value of large data analysis can be realized, and a transparent and comprehensive business view is created, thus creating the opportunity for the rapid development of the business.

First, companies should make plans to explicitly apply large data analysis to achieve business goals. Based on these objectives, the enterprise should deploy the appropriate hardware and software to meet the challenge. Deploy business intelligence solutions based on the needs of frontline staff to help them make the best decisions. With the right technical support, enterprise users and data scientists can quickly collect and analyze new data sources and explore the insights required by the business.

Myth number three: Big data only makes sense for high-tech companies like Google, Facebook and Amazon

Whether it is an internet company, Fortune 500, or a small business, it is tied to the explosive growth of big data. Data analysis has become an important business requirement, regardless of industry or enterprise size. Nowadays, it is impossible to gain real insight from business data in enterprise operation. Companies in the world's major markets are transforming the next generation of advanced analytical applications, using vast amounts of traditional data and new data in a new way to provide deeper, smarter insights. Moreover, the competitive advantage of an enterprise depends on the ability to manage and analyze all critical data in a business environment, as well as the insights that help organizations make the best decisions.

Myth Four: Data scientists and large data analysis is the 2012 it fashion

There is no doubt that big data analysis is not a fad. As the O ' Reilly Media founder Tim O ' Reilly puts it: "We are creating a fascinating new world of data-driven applications that we have created." "At the moment, data scientists have become independent careers, fighting in the forefront of shaping the new world of business, and expert data-savvy experts will become important members of the new era."

Data scientists must be curious about the data, have a focused attitude, be aggressive and be good at critical thinking. They have a deep understanding of business processes, combined with math, statistics, and skills such as Excel, SQL, and Analytical Workbench. At present, the market has a huge demand for professionals with technical competence and business awareness.

Myth Five: The value of large data depends on the technical processing power of Hadoop and similar software

No single technology can meet all requirements. Building large data analysis capabilities requires a perfect mix of talent, processes, and technologies, and the most critical is the commercial value of releasing the data, depending on the business issues that the enterprise strives to address. This will require complex analytical applications, including digital marketing optimization, fraud detection and prevention, and social network analysis.

Hadoop has some value and important position in large data technology library. Hadoop is both a framework and an excellent platform for multiple structured data filtering, transformation, and integration, similar to a sports car chassis that does not carry an engine or body. With this architecture, Hadoop can support iterative and real-time data exploration and analysis, and quickly discover patterns of change in new data and data.

The key to Success

The key to success is the ability to integrate business with traditional business data and new data. Through open access to the entire enterprise ecosystem and the integration of data from various sources, enterprises can apply large data analysis to the customer for a super comprehensive analysis, further improve customer service and sales performance. (Author: Zhang Jin Cang teradata product technology and sales support deputy general manager of Greater China region)

(Responsible editor: Lu Guang)