Performance Testing on Marker Clustering and Heatmap Visualization Techniques: A Comparative Study on JavaScript Mapping Libraries

Netek Netek,Tomecka Tomecka,Brus Brus

doi:10.3390/ijgi8080348

Netek Netek, Tomecka Tomecka + Show 1 more

Open Access

https://doi.org/10.3390/ijgi8080348

Copy DOI

Abstract

We are now generating exponentially more data from more sources than a few years ago. Big data, an already familiar term, has been generally defined as a massive volume of structured, semi-structured, and/or unstructured data, which may not be effectively managed and processed using traditional databases and software techniques. It could be problematic to visualize easily and quickly a large amount of data via an Internet platform. From this perspective, the main aim of the paper is to test point data visualization possibilities of selected JavaScript Mapping Libraries to measure their performance and ability to cope with a big amount of data. Nine datasets containing 10,000 to 3,000,000 points were generated from the Nature Conservation Database. Five libraries for marker clustering and two libraries for heatmap visualization were analyzed. Loading time and the ability to visualize large data sets were compared for each dataset and each library. The best-evaluated library was a Mapbox GL JS (Graphics Library JavaScript) with the highest overall performance. Some of the tested libraries were not able to handle the desired amount of data. In general, an amount of less than 100,000 points was indicated as the threshold for implementation without a noticeable slowdown in performance. Their usage can be a limiting factor for point data visualization in such a dynamic environment as we live nowadays.

Highlights

Big data has become a very common subject in technical, academic, and scientific publications in recent years
As a popular buzzword and objective topic of research, there are several approaches and perspectives on Big data and several ways of interpreting it that differ according to different fields of study, including geographical information science (GISci)
Many of them have open source code, typically represented by the Apache Hadoop framework, which is the most widely used technology in Big data for GISci [2]. It combines commonly-available hardware with open source software, and its development is supported by several large companies such as Google, Amazon, Microsoft, Facebook, and Twitter [3], which are looking at options for the development of internet Big data tasks

Summary

Introduction

Big data has become a very common subject in technical, academic, and scientific publications in recent years. As a popular buzzword and objective topic of research, there are several approaches and perspectives on Big data and several ways of interpreting it that differ according to different fields of study, including geographical information science (GISci). As both GISci and Big data are based on visualizations, combining the fields has potential. Several technologies used for Big data processing already exist and are being continually improved Most of these technologies are generally available as a cheap solution. Many of them have open source code, typically represented by the Apache Hadoop framework, which is the most widely used technology in Big data for GISci [2]. It combines commonly-available hardware with open source software, and its development is supported by several large companies such as Google, Amazon, Microsoft, Facebook, and Twitter [3], which are looking at options for the (future) development of internet Big data tasks

Objectives

Methods

Findings

Conclusion