The best data in the universe isn't much use if it isn't processed. Data processing refers to methods that take the raw data and turn it into usable information. Paper and pencil can work, but in the 21st century, data analysis usually relies on computers. To process data by computer, it has to be collected, checked for accuracy and entered into the computer first.
Batch processing is grunt work, the simplest form of data processing. It's useful when an organization has a large volume of data that can be clumped into one or two categories. A store, for example, can batch-process its transactions at the end of the day or the week, sending the results to the head office. If the information doesn't have to be updated for every change, batch processing is fast enough.
Sometimes batch-processing isn't fast enough. Real-time processing methods handle data when it requires an instant turn-around. If someone buys an airline ticket or cancels a reservation, for instance, the airline needs to update its records instantly. A radar system has to give its operator immediate feedback on what it detects; an ATM has to process your request for money promptly. Where batch processing handles large loads of data at specified times, real-time processing is continuous.
Data mining takes data from multiple sources and pools and combines it to look for correlations. For example, a grocery chain might analyze customers' purchases and discover that customers who buy cereal often buy bananas to go with it. The chain can use that information to increase sales, perhaps by placing bananas close to the cereal to encourage more joint purchases. The chain can also track which items sell better when the store offers coupons or holds sales.
Statistical processing involves heavy number-crunching. A company that knows it's busy on Friday can use statistical processing to calculate the effect of different variables. Some of the rush may be due to customers with last-minute requests, for instance, while another part might be a result of employees slacking off earlier in the week. Knowing the cause helps the company cope with the rush. Statistics also make it easier to compare data from different-size companies or different-size cities.
- Enterprise Features: Six Important Stages in the Data Processing Cycle
- MBA Knowledge Base: Data Processing Methods
- Data Science Central: Batch vs. Real Time Data Processing
- University of California Los Angeles: Data Mining: What Is Data Mining?
- Food and Agriculture Organization of the United Nations: Methodology
- University of Michigan: Data Processing and Statistical Adjustment