Schmidt, Benedikt (2013)
Information Work Support Based on Activity Data.
Technische Universität Darmstadt
Dissertation, Erstveröffentlichung
Kurzbeschreibung (Abstract)
In industrial and post industrial nations like Germany and the USA more than a quarter of the workforce mainly works with information. Most of the work done by these information workers is the production, supervision and dissemination of information at computer workplaces. Information workers frequently works on multiple tasks in parallel. Few guidelines regulate and structure the work process. Therefore, the successful execution of the work requires a high degree of individual planning. A common effect of ad-hoc executions of multiple tasks are memory failures: Planned activities are forgotten (prospective memory failures), or the recall of work processes' status and involved information objects fails (retrospective memory failures). The computer---a multitasking machine---even increases the likelihood of memory failures due to an increased number of activities executed in parallel. This dissertation investigates methods to decrease the likelihood of memory failures in information work at the computer workplace. The effort leads to the design of a tool that provides support for information work based on externalized activity data. This document is structured as follows: 1)The first part investigates information work from the perspectives of psychology, organization theory and sociology. Identified characteristics of information work relevant for this dissertation are captured in an ideal type. This includes the specification of the information work process at the computer workplace as being coordinated by interruptions and as being composed of logical units of work, so called knowledge actions and desktop operations. 2) The second part proposes a system design method which facilitates the analysis of work processes that can be typically observed in information work. The method seamlessly integrates into the user-centred design method. Work is modeled and analyzed in terms of so called activity system models based on activity theory and action regulation theory. System model and analysis realize two important elements of the user-centred design method: the context of use analysis and the requirement specification. The specified method is applied to the domain of information work, resulting in requirements for a tool to decrease the likelihood of memory failures in information work. 3) The third part develops methods to address memory failures in information work based on activity data. The developed methods address the requirements previously identified by applying the system design method (part 2) to the identified ideal type (part 1). The methods are implemented and evaluated in a demonstrator: a) Activity Data: A fundamental contribution to address memory failures is the collection of information about the work process. To realize this, methods to capture, analyze and organize interaction histories are developed. A core element of the process is activity mining, which is a method to identify activities in interaction histories even if the activities were interrupted during the execution process. Activity mining is modeled as a clustering problem. The proposed activity mining methods show better results than the state of the art with respect to the identification of activities. Furthermore, the proposed activity mining methods extract more details about the work execution process than the state of the art. b) Methods to Address Memory Failures: Based on the extracted activity data the goal of this work is realized---support methods to address memory failures at the computer workplace are developed. A support method design space to address memory failures is created. The design space is structured along three support directions (exploration, organization, recommendation). For each support direction, a respective user support method has been designed: 1) Activity-centric task management, which leverages activity data to facilitate task management and to support the recall of ongoing activities and respective work processes. 2) An interactive activity history, which enables the exploration of activity data in a work history visualization to support the recall of earlier work processes. 3) A recommender system, which analyzes the most recent work activities of the user to propose useful information objects like emails, files and websites. The system can be configured to support for more multitasking oriented or for more focused work. c) Transparency Tool: The support methods have been implemented in a demonstrator named Transparency. Using the demonstrator an evaluation of the support methods with a focus on memory support was conducted. The evaluation results indicate that the support methods decrease the likelihood of prospective and retrospective memory failures for information work at the computer workplace. The scientific contributions of this dissertation address two domains. On the one hand, information work support. Methods are developed which decrease the likelihood of prospective and retrospective memory failures based on activity data. On the other hand, system design methods. A method is introduced to design systems for work types which involve a high degree of individual planning.
Typ des Eintrags: | Dissertation | ||||
---|---|---|---|---|---|
Erschienen: | 2013 | ||||
Autor(en): | Schmidt, Benedikt | ||||
Art des Eintrags: | Erstveröffentlichung | ||||
Titel: | Information Work Support Based on Activity Data | ||||
Sprache: | Englisch | ||||
Referenten: | Mühlhäuser, Prof. Dr. Max ; Schmidt, Prof. Dr. Albrecht | ||||
Publikationsjahr: | 14 Mai 2013 | ||||
Datum der mündlichen Prüfung: | 27 Juni 2013 | ||||
URL / URN: | http://tuprints.ulb.tu-darmstadt.de/3652 | ||||
Kurzbeschreibung (Abstract): | In industrial and post industrial nations like Germany and the USA more than a quarter of the workforce mainly works with information. Most of the work done by these information workers is the production, supervision and dissemination of information at computer workplaces. Information workers frequently works on multiple tasks in parallel. Few guidelines regulate and structure the work process. Therefore, the successful execution of the work requires a high degree of individual planning. A common effect of ad-hoc executions of multiple tasks are memory failures: Planned activities are forgotten (prospective memory failures), or the recall of work processes' status and involved information objects fails (retrospective memory failures). The computer---a multitasking machine---even increases the likelihood of memory failures due to an increased number of activities executed in parallel. This dissertation investigates methods to decrease the likelihood of memory failures in information work at the computer workplace. The effort leads to the design of a tool that provides support for information work based on externalized activity data. This document is structured as follows: 1)The first part investigates information work from the perspectives of psychology, organization theory and sociology. Identified characteristics of information work relevant for this dissertation are captured in an ideal type. This includes the specification of the information work process at the computer workplace as being coordinated by interruptions and as being composed of logical units of work, so called knowledge actions and desktop operations. 2) The second part proposes a system design method which facilitates the analysis of work processes that can be typically observed in information work. The method seamlessly integrates into the user-centred design method. Work is modeled and analyzed in terms of so called activity system models based on activity theory and action regulation theory. System model and analysis realize two important elements of the user-centred design method: the context of use analysis and the requirement specification. The specified method is applied to the domain of information work, resulting in requirements for a tool to decrease the likelihood of memory failures in information work. 3) The third part develops methods to address memory failures in information work based on activity data. The developed methods address the requirements previously identified by applying the system design method (part 2) to the identified ideal type (part 1). The methods are implemented and evaluated in a demonstrator: a) Activity Data: A fundamental contribution to address memory failures is the collection of information about the work process. To realize this, methods to capture, analyze and organize interaction histories are developed. A core element of the process is activity mining, which is a method to identify activities in interaction histories even if the activities were interrupted during the execution process. Activity mining is modeled as a clustering problem. The proposed activity mining methods show better results than the state of the art with respect to the identification of activities. Furthermore, the proposed activity mining methods extract more details about the work execution process than the state of the art. b) Methods to Address Memory Failures: Based on the extracted activity data the goal of this work is realized---support methods to address memory failures at the computer workplace are developed. A support method design space to address memory failures is created. The design space is structured along three support directions (exploration, organization, recommendation). For each support direction, a respective user support method has been designed: 1) Activity-centric task management, which leverages activity data to facilitate task management and to support the recall of ongoing activities and respective work processes. 2) An interactive activity history, which enables the exploration of activity data in a work history visualization to support the recall of earlier work processes. 3) A recommender system, which analyzes the most recent work activities of the user to propose useful information objects like emails, files and websites. The system can be configured to support for more multitasking oriented or for more focused work. c) Transparency Tool: The support methods have been implemented in a demonstrator named Transparency. Using the demonstrator an evaluation of the support methods with a focus on memory support was conducted. The evaluation results indicate that the support methods decrease the likelihood of prospective and retrospective memory failures for information work at the computer workplace. The scientific contributions of this dissertation address two domains. On the one hand, information work support. Methods are developed which decrease the likelihood of prospective and retrospective memory failures based on activity data. On the other hand, system design methods. A method is introduced to design systems for work types which involve a high degree of individual planning. |
||||
Alternatives oder übersetztes Abstract: |
|
||||
Freie Schlagworte: | information work, context, user support, interaction history, activity mining, System design, activity theory, action regulation theory | ||||
URN: | urn:nbn:de:tuda-tuprints-36522 | ||||
Sachgruppe der Dewey Dezimalklassifikatin (DDC): | 000 Allgemeines, Informatik, Informationswissenschaft > 004 Informatik | ||||
Fachbereich(e)/-gebiet(e): | 20 Fachbereich Informatik > Telekooperation 20 Fachbereich Informatik |
||||
Hinterlegungsdatum: | 10 Nov 2013 20:55 | ||||
Letzte Änderung: | 10 Nov 2013 20:55 | ||||
PPN: | |||||
Referenten: | Mühlhäuser, Prof. Dr. Max ; Schmidt, Prof. Dr. Albrecht | ||||
Datum der mündlichen Prüfung / Verteidigung / mdl. Prüfung: | 27 Juni 2013 | ||||
Export: | |||||
Suche nach Titel in: | TUfind oder in Google |
Frage zum Eintrag |
Optionen (nur für Redakteure)
Redaktionelle Details anzeigen |