Putting Text Analytics To Work
There are three types of people in this world. Those who are good at math and those who are not. According to an IBM Marketing Cloud study, 90% of the data on the internet has been created since 2016 and the vast majority is not numbers. Those who are not good at math are adding to the worlds corpus primarily through social media. Every minute on Facebook over half a million new comments are posted, 293,000 statuses are updated, and 136,000 photos are uploaded.
Text analytics lets the type of people who are good at math meet those who have delivered a staggering amount of words to the internet together. The last several years have demonstrated a growing adoption of software for improved decision making utilizing text analytics.
We see organizations using data more strategically to discover hidden relationships. Until recently, the vast majority of information we were able examine for clients largely numerical data housed in relational databases. Yet an estimated 75 percent or more of any organizations information is unstructured. Add to this the unstructured information about any particular company or industry online and the the volume of information that can be gleaned is staggering.
The purpose of text mining is to process this unstructured information, extract meaningful numeric indexes from the text, and then access the information in the text to the various data algorithms and machine learning. Information can be extracted to derive summaries for the words contained in the documents or to calculate summaries for the documents based on the words they contain, the emotions they are conveying, or even more sophisticated types of analysis.
What business value may be gathered if analytical powered programs may be augmented with information gleaned from free form text? While information organization is certainly an important capacity of this type of analysis, we think the real value is deeper. We see the most potential in using text analytics to aid in discovery. Algorithms and machine learning allow for uncovering trends and patterns that would be otherwise unseen. This is not very different from the way we’ve seen revolutions in data analytics before, only now we can start to allow forecasting from the incredible amount of textual information now available.
It’s an exciting time in the field of text analytics. We’re enjoying the discovery, not only of the insights for our clients, but also of the continued expansion of the methodologies and outputs that text analytics can provide.