Knowledge discovery through drive test data visualization

With the increasing number of cellular phone service subscribers, the telecommunications service providers have placed immense emphasis on improving audio quality and ensure fewer dropped calls. Handoff behavior of all handsets is an important factor in quality of service of a mobile phone service. This thesis focuses on the analysis of large volumes of diagnostic data collected from mobile phones in the real world and the identification of aberrant behavior of a mobile handset under test by means of drive test data visualization. Our target was to identify poor mobility decisions that are made by the handsets in calls. Premature, delayed or exceedingly sensitive decisions are considered poor mobility decisions. The goal was to compare a set of behaviors from a baseline unit (one accepted to generally operate well). We were able to identify a particular call that was exhibiting a different path (talking to a different cell than expected or taking longer to move to a new cell). We designed a chi-square statistical test to evaluate the performance of specific mobile handset models. We also developed a mobility tool that evaluated the handset's performance by means of mapping the handoffs on the Google Maps. The mapping of the handoffs by means of the Google Maps were very powerful in identifying the above mentioned mobility patterns.