A Descriptive Statistical Analysis of Overweight and Obesity Using Big Data

In this paper, we have obtained the dataset from an open-source repository for obese people by focused on a descriptive statistical analysis of overweight and obesity using big data. We performed the statistical analysis on large scale streaming data for obesity prediction. We have classified the obesity with all categories on the scale of Body Mass Index (BMI) is being calculated i.e., underweight, normal weight, overweight, obese, very obese, and extremely obese using MapReduce technique with the help of Apache Spark and Apache Hadoop engine in pydoop python programming. The MapReduce technique in-volves the updating of cluster centers after arrival of new batch in the stream of data. The streaming of data is produced by the sensors which are classified into six different BMI categories, which are stored and processed through big data tools connected to the statistical analysis system. The Apache spark produces the latency values in accessing the data from dataset. We analyzed any obesity in the people from the normal latency value using the Apache spark and Hadoop which are well known in big data. The methods and techniques by which we can predict obesity efficiently from the large-scale streaming data has been per-formed using python programming. This is applied with the help of Apache Spark and Hadoop. In order to validate the efficiency of MapReduce technique. We have tested it both on single and distributed environment for obesity prediction using the built-in Pydoop package in python.