Call Data Record (CDR) Analytics MapReduce Case Study

In this post, we are going to solve some real world problems with map reduce code on hadoop. In this example we are going to solve a very useful problem called Call Data Record (CDR) Analytics.

Problem Statement : 

We have to write a map reduce code to find out all phone numbers who are making more than 60 mins of STD calls. Here if STD Flag is 1 that means it was as STD Call. STD is call is call which is made outside of your state or long distance calls. By identifying such subscribers, telecom company wants to offer them STD(Long Distance) Pack which would efficient for them instead spending more money without that package.

The Dataset :

A Telecom company keeps records for its subscribers in specific format. Consider following format –

First of all we need to understand the data, depending upon the output we are expecting, we need to write a mapper class which would emit FromPhoneNumber and Duration of STD Call intermediate key value pairs. To make it simple to remember the data sequence, let’s create a constants class as shown below


package hadoopgyaan.cdr;

public class CDRConstants {

    public static int fromPhoneNumber = 0;

    public static int toPhoneNumber = 1;

    public static int callStartTime = 2;

    public static int callEndTime = 3;

    public static int STDFlag = 4;



Now, lets create the mapper class which would emit intermediate key value pairs as (FromPhoneNumber, Duration), here we would also need to use our Java skills to calculate duration ( CallEndTime- CallStartTime). We are also making some manipulations to get duration in minutes.


public static class TokenizerMapper extends
            Mapper< Object , Text, Text, LongWritable> {
           Text phoneNumber = new Text();
        LongWritable durationInMinutes = new LongWritable();
        public void map(Object key, Text value,
                Mapper< Object , Text, Text, LongWritable>.Context context)
                throws IOException, InterruptedException {
            String[] parts = value.toString().split(“[|]”);
            if (parts[CDRConstants.STDFlag].equalsIgnoreCase(“1”)) {
                String callEndTime = parts[CDRConstants.callEndTime];
                String callStartTime = parts[CDRConstants.callStartTime];
                long duration = toMillis(callEndTime) – toMillis(callStartTime);
                durationInMinutes.set(duration / (1000 * 60));
                context.write(phoneNumber, durationInMinutes);
        private long toMillis(String date) {
            SimpleDateFormat format = new SimpleDateFormat(
                    “yyyy-MM-dd HH:mm:ss”);
            Date dateFrm = null;
            try {
                dateFrm = format.parse(date);
            } catch (ParseException e) {
            return dateFrm.getTime();

Now that we have already done majority of things in Mapper Class itself, here a reduce would be a simple Sum Reducer. Here is how the code would look like.

public static class SumReducer extends
            Reducer< Text , LongWritable, Text, LongWritable> {
        private LongWritable result = new LongWritable();
        public void reduce(Text key, Iterable< LongWritable> values,
        Reducer< Text , LongWritable, Text, LongWritable>.Context context)
        throws IOException, InterruptedException {
            long sum = 0;
            for (LongWritable val : values) {
               sum += val.get();
            if (sum >= 60) {
                context.write(key, this.result);

Now we can take look at the Driver class

public static void main(String[] args) throws Exception {
        Configuration conf = new Configuration();
        if (args.length != 2) {
            System.err.println(“Usage: stdsubscriber < in> < out>”);
        Job job = new Job(conf, “STD Subscribers”);
        FileInputFormat.addInputPath(job, new Path(args[0]));
        FileOutputFormat.setOutputPath(job, new Path(args[1]));
        System.exit(job.waitForCompletion(true) ? 0 : 1);

Downloads : 

CDR Sample Dataset

I hope this tutorial will surely help you. If you have any questions or problems please let me know.

Happy Hadooping with Patrick..

One thought on “Call Data Record (CDR) Analytics MapReduce Case Study

  1. I must say you have very interesting articles
    here. Your blog can go viral. You need initial boost only.
    How to get it? Search for; Etorofer’s strategies

Leave a Reply

Your email address will not be published. Required fields are marked *