How to make "Big data" better for enterprise operation service?

The sky has fallen again. This time it was "big data" that formidable IT department. As gossip rumors, no matter where you go, the discussion of "big data" is everywhere. Search for this phrase in Google, search results more than 1.3 billion. It even has a special entry in Wikipedia. The deluge of data has led many to conclude that businesses will be overwhelmed. This is not to say that the amount of information inside the enterprise will not grow. On the contrary, the enterprise's internal information is also difficult to escape the growth of the fate. Because big data is always a problem.

Despite persistent claims that the torrent of data will lead to doom, the IT industry has always been able to improve its computing infrastructure to make them faster, more capacity, cheaper, and smaller, making the lingering message "Armageddon" self-fulfilling.

Today, by using a column-type analysis infrastructure, organizations can ignore the "big data" anxiety and, conversely, make "big data" better for business operations. In a column database, users can invoke and analyze large datasets at any time, even for large datasets of various data types, such as unstructured data. Not only are they available at any time, they are faster to perform, they can be expanded more easily according to the requirements of the job, thus serving as many users as possible to cover as much data as possible.

This practice is to tap the Organization's internal and external "large data", and extract valuable parts for the enterprise to use. Its purpose is to make the organization more flexible, more competitive and improve the profitability of the organization.

One of the most important steps in deploying a data warehouse for analysis is to find quality-qualified data. From data purification to the adoption of a general strategy for data management-the technology used to ensure data quality is ripe. Internal auditing is also needed to obtain the best quality data.

Y Data Latency: Consider three aspects of data latency within your organization: Timing of data occurrence, duration of events, and time required for decision making.

Y Data Association: Work with business users to determine the relationship of data and to establish interlinkages with multiple datasets in use, while also considering data growth rates and duplication of sources.

Y self-service: Determines how advanced users can control the data used for queries without affecting it or other resources.

Y chief Data Officer: Designate a senior staff member as chief data officer to enable it to maintain organizational governance while ensuring data operability.

The importance of data quality cannot be overemphasized. In the case of comscore, a cloud company that provides analytics services and solutions for the E-commerce market, the company has learned from its inception that the focus of online marketing is shifting from the number of visitors to profitability. ComScore's client knowledge platform (Customer knowledge Platform) provides a holistic view of the behavior and preferences of customers browsing the Internet. The service tracks all users who are willing to provide Internet behavior for analysis, recording their surfing and buying behavior on various websites.

With millions of web users registering for the service to be monitored, comscore collects a huge amount of data. In fact, ComScore's analysis of compressed data reaches more than TB, adding nearly to GB every week. What's impressive is that, despite the sheer volume of data, you don't need time to wait for the results of your query. For these reasons, "we have been able to dig more quickly and provide results to our customers," said Ric Elert, vice president of the Engineering Division. This helps them improve marketing efficiency and develop more business. ”

In addition, the company uses column storage technology to achieve a 40% compression rate. Using traditional methods, the cost of storage would be much higher than it is today, ComScore said. "Because we're dealing with massive amounts of data, compression is critical to us," said Scott Smith, vice president of data warehousing. The amount of data we have is so vast that most people have never seen it. ”

Spain Airtel Vodafone's column storage Data Warehouse can be organized according to the company's business map. Although many different departments use the same data, Airtel Vodafone can still effectively guarantee the consistency and completeness of the information. The Data warehouse transforms the data into knowledge, and through an interface, converts facts in the real world into valuable business intelligence. The ability to accurately analyze and forecast customer behavior is key to Airtel Vodafone's overall business strategy.

With a column-type data warehouse, users need to get information based on workflows (rather than the hierarchy of the enterprise), which improves employee productivity and effectiveness. In other words, marketing users are using the same information as those who do financial work (for example), but they have different angles of contact with data and different analytical purposes. The Data Warehouse environment includes marketing database, call system, customer service, Global mobile communication system Statistic data, invoicing system, collection and retrieval, and all logistic management information.

Today, Airtel Vodafone has an ideal operating environment to meet a variety of requirements, enabling fast, low-cost integration of data stored in a variety of operational environments. Therefore, it can invoke detailed or summary information about company activities directly from the Data Warehouse platform. The data warehouse based on column storage makes Airtel Vodafone win market share and become a vassal of the European telecom industry.

Today, the analytics industry has no excuse for not using "big data". Whether it's expanding analytical data warehouses, covering thousands of users, or analyzing data from a variety of exotic sources, such as massive unstructured information from social media sites, they have no excuse to escape. No more hiding, the analysis industry can no longer hide behind the scary monster "big data", because we know that by using a column analysis infrastructure, we can make "big data" better for business operations.

(Responsible editor: The good of the Legacy)