Interactive exploratory data analysis using web technologies

OData support
Dr. Kocsis Imre
Department of Measurement and Information Systems

In engineering it was always important to explore and understand problems as well as possible to help find the best solution. To achieve this, a very straightforward method is to observe and measure different phenomena and then analyze the results. One available method is called Exploratory Data Analysis (EDA) which aims to discover the characteristics of variables and find correlations between them. Its main tool is to present data as graphical images and allow different interactions on them, which works well with small datasets only. However measurement tools nowadays are able to produce huge amounts of data in a very short time. My goal was to examine the possibilities of extending the common EDA ways to be able to support Big Data exploration. I also took into consideration today’s trends and found out that cloud technology holds great potential handling Big Data problems. The cloud however cannot be separated from the web which makes a necessity for my possible solutions to be available online through the internet.

In this paper I will implement an interactive EDA tool running in the browser, which will be able to support Big Data discoveries. I will investigate the various issues that arise when working with Big Data and examine the possible solutions as well. As a proof of concept I will implement a precalculation based solution to visualize Big Data, which I will test on 2 different platforms.


Please sign in to download the files of this thesis.