Real-Time Analysis of Online Sources for Supporting Business Intelligence Illustrated with Bitcoin Investments and IoT Smart-Meter Sensors in Smart Cities

Iván García-Magariño,Moustafa M Nasralla,Shah Nazir

doi:10.3390/electronics9071101

Abstract

Real-time data management analytics involve capturing data in real-time and, at the same time, processing data in a light way to provide an effective real-time support. Real-time data management analytics are key for supporting decisions of business intelligence. The proposed approach covers all these phases by (a) monitoring online information from websites with Selenium-based software and incrementally conforming a database, and (b) incrementally updating summarized information to support real-time decisions. We have illustrated this approach for the investor–company field with the particular fields of Bitcoin cryptocurrency and Internet-of-Things (IoT) smart-meter sensors in smart cities. The results of 40 simulations on historic data showed that one of the proposed investor strategies achieved 7.96% of profits on average in less than two weeks. However, these simulations and other simulations of up to 69 days showed that the benefits were highly variable in these two sets of simulations (respective standard deviations were 24.6% and 19.2%).

Highlights

Real-time data analytics usually involve both obtaining streams of live data and analyzing all data consistently [1]
We applied the proposed approach with human-centric artificial intelligence (HAI) auto-generated explanations in this context to suggest the best dates to apply the different business decisions, based on the variations of moving means (MMs) of energy consumption. This approach is illustrated in the domains of cryptocurrency investment and energy consumption in smart cities
The proposed framework uses a common notation for extracting online information with specific attributes of the RealTimeTarget class

Summary

Introduction

Real-time data analytics usually involve both obtaining streams of live data and analyzing all data consistently [1]. This kind of analytics has been widely used for business intelligence in a wide range of applications. Hammou et al [2] proposed a real-time processing framework for analyzing social big data based on a distributed recurrent neural network. This framework used deep learning techniques for supporting decision-making processes based on SMA. SMA is usually related with customer engagement and business performance [3]

Objectives

Methods

Results

Discussion

Conclusion