Good question! And I don’t have an answer for that. It’s very difficult to determine how big of the data should be called “Big data”. That’s why people keep adding more Vs characteristics to the term “Big data”. One of my project, we applied a machine learning technique (suppose to be a technique for big data analysis) to only 400 patient records (with many variables), but it turns out well. However, the more data we have, the more precision and accuracy of data analysis would be.