A data model based approach for visual analytics of monitoring data

Research output: ThesisLicenciate

Abstract

Data modelling is a well-established method in software engineering. This work explores its use in the emerging field of visual analytics. Visual analytics is a recent approach to finding knowledge from data masses. It combines the strengths of automatic data processing and the visual perception and analysis capabilities of the human user. The approach has its roots in information visualization and data analysis, in which the use of data models is not common practice. The backbone of this work is the domain data model. The model incorporates the main concepts of a given domain, which remain similar regardless of the application, but which can be tuned for visualization and analysis purposes. The work proposes three uses for data models. The first is the construction of visual analytics applications in the domain. The second is supporting reasoning with the help of metadata. The third is using the data model as an approach to visualize large data spaces. The study focuses on the analysis of monitoring data, which is nowadays collected in vast amounts and from a wide variety of fields. The approach is evaluated using two cases from different applications in the monitoring data domain: analysing the eating and exercise habits of dieting people, and studying the energy efficiency and indoor conditions of buildings. In addition to the approach and the evaluation cases, the work introduces visual analytics, data modelling and monitoring data, and discusses the evaluation of visual analytics. The multi-discipline research area of visual analytics is represented in the form of a framework constructed as a part of this work. The results suggest that data modelling is a useful method in visual analytics. A domain model approach can save effort in constructing new visual analytics applications. Supporting reasoning and browsing data with the help of the data model would be especially useful for users who are not so familiar with data analysis, but know the application domain well. Combining the data model approach with descriptive visualizations can bring powerful tools for analysing data.
Original languageEnglish
QualificationLicentiate Degree
Awarding Institution
  • Aalto University
Supervisors/Advisors
  • Puolamäki, Kai, Supervisor, External person
Publisher
Publication statusPublished - 2013
MoE publication typeG3 Licentiate thesis

Fingerprint

Data structures
Monitoring
Visualization
Metadata
Energy efficiency
Software engineering

Keywords

  • visual analytics
  • information visualization
  • data modelling
  • monitoring data

Cite this

@phdthesis{a766e1b7be5f4f01951362f3932b75aa,
title = "A data model based approach for visual analytics of monitoring data",
abstract = "Data modelling is a well-established method in software engineering. This work explores its use in the emerging field of visual analytics. Visual analytics is a recent approach to finding knowledge from data masses. It combines the strengths of automatic data processing and the visual perception and analysis capabilities of the human user. The approach has its roots in information visualization and data analysis, in which the use of data models is not common practice. The backbone of this work is the domain data model. The model incorporates the main concepts of a given domain, which remain similar regardless of the application, but which can be tuned for visualization and analysis purposes. The work proposes three uses for data models. The first is the construction of visual analytics applications in the domain. The second is supporting reasoning with the help of metadata. The third is using the data model as an approach to visualize large data spaces. The study focuses on the analysis of monitoring data, which is nowadays collected in vast amounts and from a wide variety of fields. The approach is evaluated using two cases from different applications in the monitoring data domain: analysing the eating and exercise habits of dieting people, and studying the energy efficiency and indoor conditions of buildings. In addition to the approach and the evaluation cases, the work introduces visual analytics, data modelling and monitoring data, and discusses the evaluation of visual analytics. The multi-discipline research area of visual analytics is represented in the form of a framework constructed as a part of this work. The results suggest that data modelling is a useful method in visual analytics. A domain model approach can save effort in constructing new visual analytics applications. Supporting reasoning and browsing data with the help of the data model would be especially useful for users who are not so familiar with data analysis, but know the application domain well. Combining the data model approach with descriptive visualizations can bring powerful tools for analysing data.",
keywords = "visual analytics, information visualization, data modelling, monitoring data",
author = "Paula J{\"a}rvinen",
year = "2013",
language = "English",
publisher = "Aalto University",
address = "Finland",
school = "Aalto University",

}

A data model based approach for visual analytics of monitoring data. / Järvinen, Paula.

Aalto University, 2013.

Research output: ThesisLicenciate

TY - THES

T1 - A data model based approach for visual analytics of monitoring data

AU - Järvinen, Paula

PY - 2013

Y1 - 2013

N2 - Data modelling is a well-established method in software engineering. This work explores its use in the emerging field of visual analytics. Visual analytics is a recent approach to finding knowledge from data masses. It combines the strengths of automatic data processing and the visual perception and analysis capabilities of the human user. The approach has its roots in information visualization and data analysis, in which the use of data models is not common practice. The backbone of this work is the domain data model. The model incorporates the main concepts of a given domain, which remain similar regardless of the application, but which can be tuned for visualization and analysis purposes. The work proposes three uses for data models. The first is the construction of visual analytics applications in the domain. The second is supporting reasoning with the help of metadata. The third is using the data model as an approach to visualize large data spaces. The study focuses on the analysis of monitoring data, which is nowadays collected in vast amounts and from a wide variety of fields. The approach is evaluated using two cases from different applications in the monitoring data domain: analysing the eating and exercise habits of dieting people, and studying the energy efficiency and indoor conditions of buildings. In addition to the approach and the evaluation cases, the work introduces visual analytics, data modelling and monitoring data, and discusses the evaluation of visual analytics. The multi-discipline research area of visual analytics is represented in the form of a framework constructed as a part of this work. The results suggest that data modelling is a useful method in visual analytics. A domain model approach can save effort in constructing new visual analytics applications. Supporting reasoning and browsing data with the help of the data model would be especially useful for users who are not so familiar with data analysis, but know the application domain well. Combining the data model approach with descriptive visualizations can bring powerful tools for analysing data.

AB - Data modelling is a well-established method in software engineering. This work explores its use in the emerging field of visual analytics. Visual analytics is a recent approach to finding knowledge from data masses. It combines the strengths of automatic data processing and the visual perception and analysis capabilities of the human user. The approach has its roots in information visualization and data analysis, in which the use of data models is not common practice. The backbone of this work is the domain data model. The model incorporates the main concepts of a given domain, which remain similar regardless of the application, but which can be tuned for visualization and analysis purposes. The work proposes three uses for data models. The first is the construction of visual analytics applications in the domain. The second is supporting reasoning with the help of metadata. The third is using the data model as an approach to visualize large data spaces. The study focuses on the analysis of monitoring data, which is nowadays collected in vast amounts and from a wide variety of fields. The approach is evaluated using two cases from different applications in the monitoring data domain: analysing the eating and exercise habits of dieting people, and studying the energy efficiency and indoor conditions of buildings. In addition to the approach and the evaluation cases, the work introduces visual analytics, data modelling and monitoring data, and discusses the evaluation of visual analytics. The multi-discipline research area of visual analytics is represented in the form of a framework constructed as a part of this work. The results suggest that data modelling is a useful method in visual analytics. A domain model approach can save effort in constructing new visual analytics applications. Supporting reasoning and browsing data with the help of the data model would be especially useful for users who are not so familiar with data analysis, but know the application domain well. Combining the data model approach with descriptive visualizations can bring powerful tools for analysing data.

KW - visual analytics

KW - information visualization

KW - data modelling

KW - monitoring data

UR - http://urn.fi/URN:NBN:fi:aalto-201305163100

M3 - Licenciate

PB - Aalto University

ER -