Predictable patterns in microtext as seen in educational applications using MXit in South Africa

Abstract

In South Africa, a number of mobile chat systems are used including Google Chat, BBM (Blackberry Messenger) and Watsapp. However, the mobile chat system which is the most widely used in South Africa is MXit which boasts tens of millions of users in a country of approximately fifty million people. In order to analyze MXit lingo for educational purposes, it was necessary to first determine whether or not there were any predictable patterns when people chatted using MXit lingo. This paper presents letter rankings, letter frequencies, word rankings, word frequencies and message length measurements in three different collections of messages and finds the measurements of similarity to be statistically significant. By investigating whether or not there are any predictable patterns in MXit lingo, a foundation is built for future research into MXit based communication and, therefore, other microtext communication systems.