Call Drop Analysis for Major Nordic Company

All mobile operators will have call records. Every day, all the calls that are made by people are recorded and a log file is maintained for these records. This is known as the CDR data. In this project, to get real insights on the data - you have to analyse over 100 million calls in the logs(each day). The objective of this analysis is to figure out how to resolve the call-drop issue. For example, if somebody is experiencing more than 17 call-drops in a month, then there is 90% chance that the person will drop out of the network. We will need to perform the call-drop analysis on the call log data, on a daily basis, so that- • We figure out who are the customers who are facing this difficulty. • Where they are located. • What is the reason for the call-drops. • Identify the customers who are at risk of getting dropped out. This will allow to advise the tech support team to optimize the towers from which more call-drops are occuring, to expand the capacity of these towers from where there are call drops occuring. To do this per day data should be analysed. Per day the data produced for one mobile network is 100 million call records. Since it cannot be fit into one machine - we will need to split that into 2 million blocks - so that all of it can fit into one machine for simultaneous analysis.