What is exploratory data analysis?
Developed by John Tukey, Exploratory data analysis (EDA) is an approach to data science emphasizing visualizations to uncover and summarize broad themes locked in a dataset. Exploratory data analysis can be especially interesting when applied to text mining. Effective text mining is bound by two problems, preparing the data and asking the right questions. Often, business people approach text mining as a method to test hypotheses. Take this real world example: an airport wants to know why it’s experiencing a sustained drop in non-aeronautical revenue, like restaurant sales and duty-free shopping. They assume the culprit is vendor pricing or location. The airport mines its social media feeds. Using text mining for EDA on this data set emphasizes a different conclusion, one the airport didn’t anticipate. The lack of free WiFi and charging stations in the main concourse mean customers are staying in the terminals, near outlets and WiFi hotspots.
What’s the best tool for exploratory data analysis?
Lexalytics AI Assembler solves both of the problems associated with text mining for exploratory data analysis. AI Assembler takes most of the work out of preparing a data set and fitting models. The approach leads to efficiency and greater percentages of precision. A data professional—like the one who works at the airport mentioned above—may now run use these models to precisely extract insights from the text data. This, too, happens in AI Assembler. Exploratory data analysis tools, like Lexalytics’ AI Assembler, allow the data to speak for itself, often through easy to read visualizations. EDA fundamental to the growth of any business.