Spring 2017/2018

Identification of user confusion in a web application

Date: March 2018
Experiment conductor: Michal Hucko, Matej Války
Supervisor: Prof. Ing. Mária Bieliková, PhD.
Short description:
Confusion on the web is a problem connected with web applications. Sometimes apps do not have to be complex to confuse users. Common problems may have newcomers, when accessing the app for the very first time. Ability to identify confusion can help us to give hints to users. These hints can help user to fulfil their needs. Hints can be combined to web application guides.
In our work we identify and predict confusion on web application. We will make a method for this purpose, using logs from the server. Except them we concentrate on user  mouse movements, clicks, scrolling and browsing patterns which describe behaviour. We consider usage of eye-tracker data. Next we apply machine learning methods on gathered data. To confirm our method we will establish an experiment, where we will collect feedback from users. At the moment we are working on selection of appropriate way to collect these feedback. These will be used as labeled data. Our solution can be combined with production systems providing web application guides.
Link to a formal description of the UX experiment: experiment description (In Slovak)


Scanpath Visualization

Date: April 2018
Experiment conductor: Ján Vnenčák
Supervisor: Ing. Róbert Móro, PhD.
Short description:
We use a variety of software products every day, so it is important to make these products intuitive and easy to use. This is one of the reasons why eye tracking has become more popular in recent years. With eye tracking we can easily find where and how the user is looking at visual stimulus. A motivation is to search for patterns or similarities in the recorded data that allow us to identify different user interaction strategies with the interface. One of the methods to find these patterns or similarities is the visualization of eye tracking data.
In our work, we analyzed existing methods of visualization of eye tracking data. We aim to find ones that are capable of supporting specific tasks in a better way then methods already implemented in commercial eye tracking software. Based on our analysis we implemented circular heat map transition diagram visualization method. The result of our work is a tool, which use this technique to visualize eye tracking data. The tool provide adjustment of the displayed data on user’s demand. We want to verify how this method can help in exploratory analysis of eye tracking data.
Link to a formal description of the UX experiment: experiment description (In Slovak)


The impact of images on user`s behavior on Web

Date: March 2018
Experiment conductor: Andrej Zaťko
Supervisor: Ing. Peter Gašpar
Short description:
In our Bachelor thesis we study the potential of the image. There is a great number of information on the Web which is in the form of an image. Especially in online stores, images enrich products with visual aspect, which is crucial for users during the decision. We observe the behavior of the user while displaying text and image information. Recommender systems attempt to suggest items which attract the user. Nowadays, for creating recommendations, mostly text information is used. However, in some cases they can’t describe the properties of the item accurately. Inclusion of images into the output of recommender’s systems seems to be the way to overcome this issue and may lead to an improvement in accuracy of recommendations.  We propose a method for the evaluation of changes in user`s behavior.
Link to a formal description of the UX experiment: experiment description (In Slovak)


Eye tracking the user behavior on the Web

Date: March 2018
Experiment conductor: Martin Svrček
Supervisor: Prof. Ing. Mária Bieliková, PhD.
Short description:
User behavior on the Web is affected by many various effects. Every person behave differently in different situations according their personal traits or states. However, there is also Web itself and especially the usability of the Web, which impact this behavior. Usability is therefore a subject of many studies and usability testing. There are many ways how to measure the usability. Often we are trying to reveal the visual attention of web users to understand their behavior. Nowadays, one of the best methods to reveal the visual attention is by tracking the user eye movement. Eye tracking is very interesting approach, which can help us to evaluate the overall usability of the Web. However, there are many different methods of analyzing the data from eye tracking and simultaneously there is no sufficient research in eye tracking data analysis in area of the user behavior on the Web. Reason for this is that Web is a dynamic environment and most of the actual research is directed only to static inputs like images. Our goal is to analyze the different representations of eye tracking data and different methods of analysis itself in order to determine the impact of usability on user behavior on the web.
Link to a formal description of the UX experiment: experiment description (In Slovak)


Identifying users characteristics by eye tracking analysis

Date: March 2018
Experiment conductor: Mária Dragúňová, Matej Červenka
Supervisor: Mgr. Jozef Tvarožek, PhD.
Short description:
Human individuality reflects on many aspects of our everyday lives and using web applications is not an exception. Problem of variablity between different users and their needs is commonly solved by personalisation to achieve the best possible usability for every user. However, human is a complex being and there are many characteristic, which should be considered when a system adaptation comes to mind. In our work, we analyze some of cognitive characteristics – working memory capacity, visual search and verbal-imagery dimension of cognitive styles. We propose to consider these characteristics when searching in e-catalog, which is a common structure in e-shop applications. Besides analysis of characteristics, we also focus on identifying tasks in e-commerce environment using eye-tracker.
Link to a formal description of the UX experiment: experiment description (In Slovak)


Identification of usability problems on web sites using the eye-tracker

Date: March 2018
Experiment conductor: Martin Šidlo
Supervisor: Ing. Eduard Kuric, PhD.
Short description: Lately, usability is more discussed in the field of information technologies. Creators of the systems realized, that user experience is a huge part of a product and has a significant impact on the usage. The most effective way, how to evaluate websites, is to use user testing methods. Evaluating of such methods is a long-lasting process. Goal of this work is to shorten this time and suggest potential usability problems without a need of evaluating each user separately. In our work, we want to create models to predict such problems on a website, based on its components. In order to be able to build it, we need to train them. For this purpose, we have to gain enough data. We are using approach of a previous work of Pang et al. [1]. They have been directing user visual flow on a website. We want to use this approach with some modification and user labeling according to their cognitive abilities in order to automatically identify usability problems on websites.
Link to a formal description of the UX experiment: experiment description (In Slovak)


Detection of User’s Intent on the Web based on Eye-tracking Method

Date: March 2018
Experiment conductor: Ondrej Kipila
Supervisor: Ing. Jakub Šimko, PhD.
Short description:
The characteristics of the user are are very useful information in information systems. If the information system is able to automatically detect the user’s characteristics (eg. mood, interest, fatigue, personality characteristics, capacity short-term memory), can use this information to adapt its functionality in such a way that as far as possible meet the needs of the user. An example would be automatic content recommendation. Another option is the use of prediction of his other activities and results in the system.
One way of identifying a characteristic of a user uses the tracking of user activity in the system (eg. Clicks, scrolling, text entry). Records of these activities can be further combined with monitoring a user’s perspective in the implementation of these activities. Other data usable to create a user model is the size of the pupil.
In this master’s thesis we deal with the prediction of the user’s intention (goal) when working with a web page, namely when working with an online shop. The aim of this thesis is to create a predictive model that analyzes the behavior and the user’s sight and predicts the intent in real time.
Link to a formal description of the UX experiment: experiment description (In Slovak)


User experience with web browsing, and the usage of control devices

Date: February 2018
Experiment conductor: Ján Trimel
Supervisor: Ing. Peter Krátky
Short description: In this experiment we are going to develop a method of estimating user’s skill in web browsing using data from monitoring the mouse and the keyboard. We determine the user’s level of experience using questionnaires. Then we find a common pattern for experienced and novice users from their mouse, keyboard usage and visual fixation.
Link to a formal description of the UX experiment: experiment description (In Slovak)


Analysis of influence  of different types of dataset to express the intensity of emotion

Date: February 2018
Experiment conductor: Adam Strásky
Supervisor: Ing. Róbert Móro, PhD.
Short description:
When we recognize emotions using a webcam, we often look at emotions only during usability testing. We often do not recognize the extent to which individuals express their emotions and how these tools are accurate because we can not determine plausibility on the basis of comparison with the basic expression of emotions. The goal of the experiment is to obtain a range of emotions with varying degrees of differentiation and to determine if during the test the emotions will be captured to correspond with the expected results.
Experiment design consists of an initialization dataset, where the stimulus will consist of images, music clips and videos. We will try to create a stimulus scale for this participant and then perform a usability test. Finally, we will use retrospective verbal protocol with each participant to refine the results.
Link to a formal description of the UX experiment: experiment description (In Slovak)


Eye Tracking Using Deep Neural Networks

Date: February 2018
Experiment conductor: Patrik Gajdošík
Supervisor: doc. Ing. Vanda Benešová, PhD.
Short description:
Eye tracking offers valuable insight in the fields of usability and user experience. However, in order to perform a study where the users’ gaze is being recorded, specialized hardware, eye trackers, are needed, which limits the span of such studies and renders them expensive and time consuming. In order to solve this issue, recent studies focus on using the ordinary web-cameras that are available on most of the today’s mobile devices. Some papers, such as ours, employ convolutional neural networks. The problem is that the variance of images shown to these networks during training is very high and we also cannot easily perform calibration as in the majority of eye tracking solutions. In our work, we focus on exploring the ways in which to alleviate the task for the neural network. We train and evaluate our models on existing datasets as well as on data gathered by proper eye tracker.
Link to a formal description of the UX experiment: experiment description (In Slovak)