Big data use cases form the framework that determines the specific insights you are seeking from any big data project, including the data sources most likely to deliver those insights, and the analytics necessary to yield the appropriate results. Big data use cases can answer questions about sales, marketing, customer preferences, operations, or anything else wherevolumes of in-house and external data can be analyzed to reveal meaningful trends.
By definition, big data analyzes both external and internal data resources to answer a specified question. A use case is a modeling technique that defines how forces interact with a system to achieve desired goals. In big data use cases, you assemble data sources and look for patterns that will provide answers to a specific business problem. What makes big data use cases so valuable is their ability to assimilate structured data, such as last quarter’s sales figures, with unstructured data, such as social media commentary, to identify trends.
Five big data use cases tend to generate the most big data value:
Big data exploration – The most common use cases are big data exploration. These use cases take massive amounts of available information from internal data sources, enterprise resources, and external resources to identify patterns that aid decision-making. The idea is to create a unified view of available data to reveal new insights.
Gaining a 360-degree view of the customer – The number of touch points between any organization and its customers has increased with the digital revolution. Big data allows you to consolidate those touch points to achieve a three-dimensional portrait of the customer. You can develop big data use cases that apply a holistic approach to identifying the customer, using past purchasing history, customer loyalty information, social media interaction, and other data sources to improve customer engagement and promote customer loyalty. Any business can benefit from customer big data use cases, especially retailers.
Operations analytics – The Internet of Things (IoT) is making it possible to tap into devices with built-in sensors, which is opening up a wide range of new analytics possibilities. Tapping data gathered from these remote sensors can assess various types of processes, from manufacturing machinery to city traffic flow. By gaining a real-time snapshot of operations, big data use cases can be developed to assess potential improvements in processes and procedures to increase efficiency and reduce operations overhead.
Security – Big data use cases look for patterns based on data input, including anomalies. That makes big data ideal for enterprise security applications. Real time big data analytics can detect abnormal users, hacking attacks, fraudulent activities, and cyber-attacks. Using real-time analytics can detect an attack and take pre-programmed action before the network is affected.
Data warehouse augmentation – This is a form of data warehouse modernization where you can improve efficiencies by identifying unstructured data sources that could benefit from Hadoop analytics. There are three basic types of augmentation:
- Pre-processing, which uses big data as a staging area to determine what data should be moved to the data warehouse;
- Offloading, which moves infrequently used data to Hadoop storage; and
- Exploration, which uses big data to find new value from raw data, and frees up the data warehouse for deeper, more structured analytics.
Once you determine the big data use cases that will promote the most value you can determine what type of analytics are best suited to the use case. Here are three basic kinds of analytics:
Descriptive Analytics assess past performance to predict likely future outcomes. Estimates are that about 80 percent of business analytics are considered descriptive, and can be applied to sales, production, shipping, or other operations. In the case of big data, you want to analyze both synchronous and asynchronous data sets to reveal new historic patterns you wouldn’t get from the raw data. Those historic patterns should shed some light on future outcomes.
Predictive Analytics allow you to make an educated guess about the future using statistical modeling and data mining. The analytics deliver degrees of probability based on the relationship between historical data sets from CRM, ERP, POS, etc. Predictive analytics can help guide purchasing patterns, supply chain and inventory requirements, sales trends, and other business factors.
Prescriptive Analytics prescribe potential actions that to quantify the impact on future decisions. These are “what if” scenarios that assess potential outcomes based on specific actions. Using historical data, read-time data feeds, business rules, and other sources predictive analytics provide recommendations based on available data. These are the most complex kinds of business analytics and potentially the most powerful.
If you understand how to formulate big data use cases and assemble the data sources necessary to answer those questions, then you will have a firm foundation on which to build a successful big data program.