For many companies today, it is highly valuable to collect and analyse data in order to support decision making and functions of various sorts. However, this kind of data-driven approach is seldomly applied to software testing and there is often a lack of verification that the testing performed is relevant to how the system under test is used. Therefore, the aim of this thesis is to investigate the possibility of introducing a data-driven approach to test automation by extracting user behaviour data and curating it to form input for testing. A prestudy was initially conducted in order to collect and assess different data sources for augmenting the testing. After suitable data sources were identified, the required data, including data about user activity in the system, was extracted. This data was then processed and three prototypes where built on top of this data. The first prototype augments the model-based testing by automatically creating models of the most common user behaviour by utilising data mining algorithms. The second prototype tests the most frequent occurring client actions. The last prototype visualises which features of the system are not covered by automated regression testing. The data extracted and analysed in this thesis facilitates the understanding of the behaviour of the users in the system under test. The three prototypes implemented with this data as their foundation can be used to assist other testing methods by visualising test coverage and executing regression tests.

No description available.

The thesis deals with the problematics of interactive books and their influence on the education of preschool children, which is in the Czech Republic still one of the unexplored areas. The aim of the thesis is to compare the paper books with the interactive books in terms of their impact on the child's learning and to see if interactivity influences the better memorization of information than interaction with the kindergarten teacher and also if the children are able to work with the interactive book. There are two hypotheses H1: A child of pre-school age working with an interactive book can remember less information than a child listening to a narrative from a classical book, and H2: A pre-school child who working with an interactive book will remember better the visual rendering of the content. The examined group was composed of preschool children from two differently different kindergartens. For the research, the method of participated target group observation was used when working with interactive books, supplemented by interviews with nursery teachers who participated in the research. The development of the pre-school child's cognitive functions and the development of its interaction with the technologies is briefly described in the theoretical part one. The technologies and their role in the...

In today’s business climate permeated by Big Data, an opportunity to drive performance lies in analysing consumer behaviour from user data. In particular for online content providers, user data is available in abundance and logged continuously. This leads to new possibilities for design and usage of metrics, as businesses can benefit from smart and timely decision-making. However, in order to profit from user data in performance measurement (PM), it is critical to identify metrics that truly guide decisions. Thus, an effective and efficient PM process is imperative. Despite its promise, Big Data’s role in PM has been scarcely researched. Research has studied user behaviour from data, for instance in the context of video or audio streaming and web search, but primarily with a focus on technical performance. In addition, the research on online content providers’ PM is fragmented, and has mainly been conducted by practitioners. Thus, the PM field needs to be updated to reflect today’s dynamic and digital business climate. Therefore, the purpose of this research was to explore how online content providers, generating a large amount of user data, work with PM, and also practically illustrate how metrics can be designed from user data. The research was carried out as a case study at an audio streaming company, but empirics was also gathered from other online content providers with the aim to increase the generalisability. The illustration of metric design was based on quantitative analysis of commuters’ in-car audio streaming. For commuters’ audio streaming it was found that suitable metrics should capture the habitual nature. Therefore engagement metrics were found to be applicable, for instance the fraction having sessions both in the morning and afternoon, and the fraction having more than one day commuting with the streaming service per week. In regard to online content providers’ PM process, this research contributes with a proposed framework, which was developed from three existing frameworks; HEART reflected as important measurement dimensions and translation of goals to metrics, OKR which sets the focus in terms of high-level goals, and design-implement-use reflected as the process’ phases. It was found that insights from user data and explicit user feedback are complementary and can arise throughout the whole process, and that mutual communication between data scientists and product managers is crucial. Further, four types of iterations were identified in the process; modifying a metric, designing new metrics, completely changing a metric, and starting new initiatives. Moreover, metrics were found to be highly context dependent. Additionally, four important aspects were identified in metric design; data availability and proxy assessment, characteristics and form of metric, metric trade-offs, and metric movement interpretation. / I dagens affärsklimat genomsyrat av Big Data finns en möjlighet att driva resultat framåt genom analys av kundbeteenden från användardata. I synnerhet för online-tjänsteföretag samlas användardata kontinuerligt och finns tillgänglig i en oerhörd mängd. Detta skapar nya möjligheter för design och användande av mätetal då företag kan utveckla smartare och snabbare beslutsfattande. För att verkligen dra fördel av användardata i prestationsmätning (PM) är det dock kritiskt att identifiera mätetal som faktiskt bistår beslutsfattande, vilket följaktligen kräver en effektiv PM-process. Trots potentialen är forskning på Big Data inom PM begränsad. Studier har analyserat kundbeteenden från användardata, exempelvis i kontexten av strömmad video eller audio och webbsökningar, men primärt med fokus på tjänstens tekniska prestanda. Vidare är forskning på PM hos online-tjänsteföretag fragmenterad, och huvudsakligen genomförd av företag inom industrin. Följaktligen bör fältet aktualiseras för att reflektera dagens digitala och dynamiska affärsklimat. Därför var syftet med denna studie att utforska hur online-tjänsteföretag, som besitter stora mängder användardata, arbetar med PM, men även praktiskt illustrera hur mätetal kan designas från denna data. Studien genomfördes som en fallstudie på ett ljud-strömningsföretag, men empiri insamlades även från andra online-tjänsteföretag med avsikt att öka generaliserbarheten. Den praktiska illustrationen av mätetals-design baserades på en kvantitativ analys av pendlares audio-strömning i bil. För pendlares audio-strömning i bil fann denna studie att lämpliga mätetal bör fånga den vanemässiga aspekten associerad med pendling. Därmed anses mätetal som reflekterar engagemang lämpliga, exempelvis andelen som har sessioner både på förmiddagen och eftermiddagen och andelen som har mer än en dag med pendlar-sessioner i veckan. Gällande PM-processen hos online-tjänsteföretag bidrar denna studie med ett föreslaget ramverk som utvecklades från tre existerande ramverk; HEART som reflekteras i form av viktiga mätetalsdimensioner samt översättning av mål till mätetal, OKR vilket sätter fokus för processen i termer av mål på högre nivå, och designa-implementera-använda som reflekterar processens faser. I studien kom det fram att insikter från användardata och explicit användaråterkoppling kompletterar varandra, och att dessa kan uppkomma under hela processen. Vidare konstaterar denna studie att ömsesidig kommunikation mellan dataforskare och produktchefer är essentiellt. Dessutom identifierades fyra typer av iterationer som kan förekomma vid användning av mätetal; modifiera mätetal, designa nya mätetal, fullständigt förändra mätetal samt påbörja nya initiativ. Därutöver kan studien konstatera att mätetal är högst kontextberoende, och att det finns fyra viktiga aspekter att ta hänsyn till i mätetals-design; data-tillgänglighet och proxy-utvärdering, karaktäristik och form på mätetal, trade-off mellan mätetal, samt tolkning av mätetals-förändringar.

Die Identifikation von Sessions zum Verständnis des Benutzerverhaltens ist ein Forschungsgebiet des Web Usage Mining. Definitionen und Konzepte werden seit über 20 Jahren diskutiert. Die Forschung zeigt, dass Session-Identifizierung kein willkürlicher Prozess sein sollte. Es gibt eine fragwürdige Tendenz zu vereinfachten mechanischen Sessions anstelle logischer Segmentierungen. Ziel der Dissertation ist es zu beweisen, wie unterschiedliche Session-Ansätze zu abweichenden Ergebnissen und Interpretationen führen. Die übergreifende Forschungsfrage lautet: Werden sich verschiedene Ansätze zur Session-Identifizierung auf Analyseergebnisse und Machine-Learning-Probleme auswirken? Ein methodischer Rahmen für die Durchführung, den Vergleich und die Evaluation von Sessions wird gegeben. Die Dissertation implementiert 135 Session-Ansätze in einem Jahr (2018) Daten einer deutschen Preisvergleichs-E-Commerce-Plattform. Die Umsetzung umfasst mechanische Konzepte, logische Konstrukte und die Kombination mehrerer Mechaniken. Es wird gezeigt, wie logische Sessions durch Embedding-Algorithmen aus Benutzersequenzen konstruiert werden: mit einem neuartigen Ansatz zur Identifizierung logischer Sessions, bei dem die thematische Nähe von Interaktionen anstelle von Suchanfragen allein verwendet wird. Alle Ansätze werden verglichen und quantitativ beschrieben sowie in drei Machine-Learning-Problemen (wie Recommendation) angewendet. Der Hauptbeitrag dieser Dissertation besteht darin, einen umfassenden Vergleich von Session-Identifikationsalgorithmen bereitzustellen. Die Arbeit bietet eine Methodik zum Implementieren, Analysieren und Evaluieren einer Auswahl von Mechaniken, die es ermöglichen, das Benutzerverhalten und die Auswirkungen von Session-Modellierung besser zu verstehen. Die Ergebnisse zeigen, dass unterschiedlich strukturierte Eingabedaten die Ergebnisse von Algorithmen oder Analysen drastisch verändern können. / The identification of sessions as a means of understanding user behaviour is a common research area of web usage mining. Different definitions and concepts have been discussed for over 20 years: Research shows that session identification is not an arbitrary task. There is a tendency towards simplistic mechanical sessions instead of more complex logical segmentations, which is questionable. This dissertation aims to prove how the nature of differing session-identification approaches leads to diverging results and interpretations. The overarching research question asks: will different session-identification approaches impact analysis and machine learning tasks? A comprehensive methodological framework for implementing, comparing and evaluating sessions is given. The dissertation provides implementation guidelines for 135 session-identification approaches utilizing a complete year (2018) of traffic data from a German price-comparison e-commerce platform. The implementation includes mechanical concepts, logical constructs and the combination of multiple methods. It shows how logical sessions were constructed from user sequences by employing embedding algorithms on interaction logs; taking a novel approach to logical session identification by utilizing topical proximity of interactions instead of search queries alone. All approaches are compared and quantitatively described. The application in three machine-learning tasks (such as recommendation) is intended to show that using different sessions as input data has a marked impact on the outcome. The main contribution of this dissertation is to provide a comprehensive comparison of session-identification algorithms. The research provides a methodology to implement, analyse and compare a wide variety of mechanics, allowing to better understand user behaviour and the effects of session modelling. The main results show that differently structured input data may drastically change the results of algorithms or analysis.