Real-Time Uber Data Analysis of Popular Uber Locations in Kubernetes Environment

摘要

Data is crucial in today’s business and technology environment. There is a growing demand for Big Data applications to extract and evaluate information, which will provide the necessary knowledge that will help us make important rational decisions. These ideas emerged at the beginning of the 21st century, and every technological giant is now exploiting Big Data technologies. Big Data refers to huge and broad data collections that can be organized or unstructured. Big Data analytics is the method of analyzing massive data sets to highlight trends and patterns. Uber is using real-time Big Data to perfect its processes, from calculating Uber’s pricing to finding the optimal positioning of taxis to maximize profits. Real-time data analysis is very challenging for the implementation because we need to process data in real-time, if we use Big Data, it is more complex than before. Implementation of real-time data analysis by Uber to identify their popular pickups would be advantageous in various ways. It will require high-performance platform to run their application. So far no research has been done on real-time analysis for identifying popular Uber locations within Big Data in a distributed environment, particularly on the Kubernetes environment. To address these issues, we have created a machine learning model with a Spark framework to identify the popular Uber locations and use this model to analyze real-time streaming Uber data and deploy this system on Google Dataproc with the different number of worker nodes with enabling Kubernetes and without Kubernetes environment. With the proposed Kubernetes environment and by increasing the worker nodes of Dataproc clusters, the performance can be significantly improved. The future development will consist of visualizing the real-time popular Uber locations on Google map.