In the past, unstructured data was too complex for common statistical programming languages to analyze. Therefore, in order to gain meaningful interpretations from unstructured data, people had to manually sort through various sources of data to draw conclusions. With limited resources and an overwhelming quantity of data, this method has become inefficient and costly. As a result, a series of tools have emerged over the years that enables analysts to process and analyze both structured and unstructured data to gain valuable insights and help guide business decisions.
More specifically, this field is known as natural language processing (NLP) or content analytics. NLP refers to techniques that allow text to be analyzed and identifies trends, patterns, and statistically relevant findings from mining thousands of data sources. Using NLP techniques, analysts can discover unexpected correlations or anomalies for businesses to investigate and can analyze at a higher level than previous technologies would allow. Businesses are increasingly using NLP to react faster to changes in data and make better business decisions. Organizations can benefit from the use of NLP in various industries such as government, insurance, healthcare, and law enforcement.
When an unexpected hurricane, tornado, or flood strikes, insurance companies can easily become exasperated with an overload of claims and as a result, desperate customers will have a negative experience with the company. It is essential for insurance companies to deliver timely and appropriate responses to their customers, especially during these hectic times. To put this into perspective, Security First Insurance settles about 700 claims in a typical month, but when there is a hurricane, they may need to settle 40,000 claims in a few days. Security First Insurance previously did not have the capability to respond promptly when contacted through email or social media channels during busy times. However, Security First Insurance is now able to evaluate incoming email and social media communication using content analytics which helps prioritize issues and deliver timely, personalized responses to customers.
NLP is particularly useful in the healthcare industry because the majority of stored health information is unstructured, making it a challenge to sort and analyze the data. For example, NLP techniques can be used to recognize medical trends and outcome predictions of patients. Further, content analytics of structured and unstructured health data can improve diagnostic assistance, research for improved disease management, patient follow up care, and readmission prevention. Seton Healthcare illustrates the value of NLP in the healthcare industry through extracting key information from unstructured health data using content analytics to identify preventative care opportunities. As a result, they were able to identify patients with congestive heart failure who were likely to be readmitted and were able to reduce readmission through discovering high risk patients.
Another industry that has proven the effectiveness of NLP is the law enforcement industry. Using NLP analytics and technology, the Police Department in Durham was able to determine incidents and individuals that were interrelated to help identify potential suspects and prevent crime. Content analytics was used to identify patterns and individuals involved in crime and gang activity, leading to the discovery that 20 percent of violent gun crime happened in a specific two square miles of Durham. These tools allowed police officers to look at density rings for “shots fired” calls, violent firearm crimes, and the last known addresses of gang members. Consequently, crime in Durham was brought down by over 50 percent because NLP allowed the police to respond to information quicker and more efficiently.