Jan Gugler, TiltablePages: Augmenting tilted papers, University of Zurich, Faculty of Business, Economics and Informatics, 2019. (Bachelor's Thesis)
Augmented reality applications are becoming increasingly commonplace in a variety of fields. However, the use of augmented reality in an academic context is limited. In this thesis, we developed a tool which allows researchers to superimpose virtual objects onto their
experiment setups. For this purpose, we have implemented an easily adaptable tracking and projection pipeline, consisting of three parts: (1) off-axis perspective projection, (2) calibration, and (3) keystone correction. This pipeline uses tracking data to compute the off-axis perspective projection, which creates a 3D illusion out of a flat surface. The resulting image is then aligned with the desired projection surface and skewed to ensure a non-distorted final image is sent to the projector. Additionally, an augmented reality computer game was developed in the Unity3D engine, to showcase the pipeline and provide a fun game-play experience. |
|
Florian Fischer, Indivi: Designing a tool for researchers to create individualized feedback, University of Zurich, Faculty of Business, Economics and Informatics, 2018. (Master's Thesis)
In the field of Psychology, researchers typically conduct studies with human participants. With the raise of new technologies (e.g., smartphones, sensors) and the way they affect humans’ everyday life, participants want to be more involved in research, help to shape research
questions and receive feedback. To fulfill this expectation, researchers in Psychology are in need of a tool that guides them and automates the creation of feedback reports for studies. This research focuses on three research questions: (1) find out the design principles for such a tool to follow the current practice of giving feedback, (2) find out the design
principles for such a tool to output individualized feedback, and (3) infer a good visualization for a specific feedback given the type of data provided. This research project elicited the requirements and developed a tool called Indivi. (1) Indivi creates individualized feedback following a workflow that was elaborated together with researchers in Psychology. (2) Indivi individualizes feedback on three levels: individualizing graphs displaying the individual value, individualizing texts by using placeholders to input the individual value, and individualizing text by classifying participants into categories and writing a text for each category. (3) Indivi also infers a fitting type of graph based on the type of data provided and some user choices. Indivi was found to represent an improvement compared to manual practices. In future developments, Indivi would benefit from being extended to a platform not only for researchers but also for participants. It could provide instant feedback generated directly from the data collection and this feedback be used to steer behavioral changes. |
|
Raphael Weibel, Annostand: Understanding annotations, University of Zurich, Faculty of Business, Economics and Informatics, 2018. (Master's Thesis)
Reading and annotating a scientific paper is an integral part of all academic fields. Differ- ent approaches for understanding annotation behavior have been proposed and tested, but none of them have identified a useful measure for discriminating the important annotations. This thesis investigates the use of reading behavior as a measure of annotation importance. A two-session experiment has been conducted. During the first session, participants were asked to read and annotate a scientific paper. During the second session, participants used their own annotations to write a summary of the article. During both sessions, participants’ eye movements were recorded. Different statistical learning methods (i.e., Support Vector Machines, Random Forest, Regression) were compared in terms of their ability to classify im- portant annotations. The results of this analysis were not able to confirm that reading behav- ior can be used to classify the importance of annotations. Future research can be conducted in order to determine whether this lack of confirmation is due to inaccurate eye movement recordings or to the inability of these measures to accurately classify important annotations. |
|
Jasmin Ebner, ActiveStat: Design Patterns and Software Platform for Interactive Learning of Statistics, University of Zurich, Faculty of Business, Economics and Informatics, 2018. (Master's Thesis)
New research areas such as Data Science and Machine Learning have increased the importance of statistics in recent years. Additionally, computers have changed the way statistics are learned. Nowadays, computers take over complex calculations and, as a result, students can focus on the underlying concepts. Due to these developments, in this thesis, we want to find out how computers and software can be used to improve statistical learning. For this purpose, we have designed and developed an application for learning statistics on the example topic linear regression. A toolbox with reusable components and design patterns was elaborated to help future developers in building such software. The evaluation of a first study with students gives evidence that the software is suitable for learning statistics and that the design patterns help to evolve students' mental models of statistical concepts. During the interaction with the software, the students drew correct conclusions and were able to describe concepts about linear regression in a posttest which they could not explain before the study (in a pretest). But the study was conducted with only a few students and, therefore, requires further evaluation in future work. |
|
Alexander Eiselmayer, Touchstone 2: A Visual Programming Interface for Designing and Comparing Experiments, University of Zurich, Faculty of Business, Economics and Informatics, 2018. (Master's Thesis)
|
|
Chia-Kai Yang, Chatchavan Wacharamanotham, Alpscarf: Augmenting Scarf Plots for Exploring Temporal Gaze Patterns, In: CHI 2018, ACM Press, New York, USA, 2018-04-21. (Conference or Workshop Paper published in Proceedings)
Scarf plots visualize gaze transitions among areas of interest (AOIs) on timelines. Nevertheless, scarf plots are ineffective when there are many AOIs. To help analysts explore long temporal patterns, we present Alpscarf, an extension of scarf plots with mountains and valleys to visualize order-conformity and revisits. Alpscarfs are rendered in two complementary modes in aid of insight discovery. An R package of Alpscarf is available at github.com/chia-kaiyang/alpscarf. |
|
Krishna Subramanian, Johannes Maas, Michael Ellers, Chatchavan Wacharamanotham, Simon Voelker, Jan Borchers, StatWire: Visual Flow-based Statistical Programming, In: Extended Abstracts of the 2018 CHI Conference on Human Factors in Computing Systems, ACM, New York, NY, USA, 2018-04-21. (Conference or Workshop Paper published in Proceedings)
Statistical analysis is a frequent task in several research fields such as HCI, Psychology, and Medicine. Performing statistical analysis using traditional textual programming languages like R is considered to have several advantages over GUI applications like SPSS. However, our examination of 40 analysis scripts written using current IDEs for R shows that such scripts are hard to understand and maintain, limiting their replication. We present StatWire, an IDE for R that closely integrates the traditional text-based editor with a visual data flow editor to better support statistical programming. A preliminary evaluation with four R users indicates that this hybrid approach could result in statistical programming that is more understandable and efficient. |
|
Chatchavan Wacharamanotham, Matthew Kay, Steve Haroz, Shion Guha, Pierre Dragicevic, Special Interest Group on Transparent Statistics Guidelines, In: Extended Abstracts of the 2018 CHI Conference on Human Factors in Computing Systems, ACM, New York, NY, USA, 2018-04-21. (Conference or Workshop Paper)
Transparent statistics is a philosophy of statistical reporting whose purpose is scientific advancement rather than persuasion. At our CHI 2017 workshop, "Moving Transparent Statistics Forward", we identified that an important first step is to develop detailed guidelines for authors and reviewers in order to help them practice and promote transparent statistics. We propose a SIG to solicit feedback from the CHI community on a first working draft of "Transparent Statistics Guidelines" and engage potential contributors to push the transparent statistics movement forward. |
|
Flavio Keller, ChartGaze: Eye Movement in Chart Reading, University of Zurich, Faculty of Business, Economics and Informatics, 2018. (Master's Thesis)
Data visualizations, especially charts, are crucial tools in research, but also in everyday life. This thesis investigates if typical eye movement sequences can be mapped to low-level tasks people usually solve when they read charts. An experiment was carried out, letting participants solve five different task types for four different chart types, totalling in 60 trials. In this experiment, eye-tracking was employed to record the movement of participants’ eyes. The recordings were used to compute features and infer Hidden Markov models. These inferred models were the basis for the analysis of the recorded data. Classification results for the utilized task types show that for many task type and chart type combinations, tasks can be predicted with an accuracy of 50% and reach up to 70%. Further investigation showed that applying the same analysis on partitions of each trial oftentimes reveal better prediction accuracies than when using the full trial. An additional comparison with baseline eye tracking features demonstrate that Hidden Markov model features in most cases obtain better results. The discussion reflects on the obtained results and analyses possible influences on the gained results and how they could be improved. Additionally, a simple implementation of a task recognizer is presented and recommendations for future studies and implementations are given. |
|
Caroline Lottenbach, Evaluating and Improving Usability of OLAT, University of Zurich, Faculty of Business, Economics and Informatics, 2017. (Bachelor's Thesis)
The goal of this bachelor thesis is to improve the usability of the online learning platform OLAT, currently used at the University of Zurich. Three different parts of OLAT version 10 were analyzed in this thesis. Namely the "Course overview" page, the "Course Editor", and the course element task types available in the course editor.
The three areas of OLAT were addressed through interviews, survey, heuristic evaluation and prototyping.
For the "Course overview" and the "Course Editor" the usability problems that were found are documented. Possible solutions to some of the problems of the "Course Editor" are provided in the form of prototypes and the differences of the task type course elements are elucidated. |
|
Silas Weber, Eye tracking Analytics in the browser, University of Zurich, Faculty of Business, Economics and Informatics, 2017. (Bachelor's Thesis)
Eye tracking analysis software can be a powerful tool to research how humans perceive and interact a with a variety of things, be it from how we perceive advertisements to how we interact with a device via its user interface.
The idea for developing a new application for analyzing the gaze points of a person came from the need to provide an interaction design that enables the user of the system to efficiently switch between the objects the gaze points fall upon. As an example, to analyze how a person reads a newspaper, the user of the application needs to be able switch between the pages at least as fast as the person who reads the newspaper flips through the pages.
This paper first gives more detailed explanation for the motivation behind developing this new tool and short introduction to the relevant theory behind analyzing eye data. The rest of the paper explains the design process and implementation detail of the new tool.
Through the power of JavaScript on the client and its capabilities on the server, this new application brings eye data analysis to the browser. Having an interaction design tailored to the needs of actual researchers, it is equipped with a player to show the fixations on the object, a drawing editor to define AOIs (areas of interest), and the functionality to create visualizations of the analysis in the form of bar charts. |
|
Till Salinger, AnnoTracer: A Tool for Annotations Extraction and Analysis, University of Zurich, Faculty of Business, Economics and Informatics, 2017. (Master's Thesis)
AnnoTracer is an application that can extract hand-written annotations from research papers and overlay them on top of the original PDF. It can also record and analyze eye tracking data in relation to the original PDF and the currently displayed annotations.
The goal of this thesis was to learn more about the influence of hand-written annotations on reading behavior and text comprehension. A study with 12 participants has been conducted and the evaluation indicates that students who use AnnoTracer might have a higher chance to notice contradictions within a research paper. However, since the sample size was very small, more research is required to better understand the underlying relationship. |
|
Matthew Kay, Steve Haroz, Shion Guha, Pierre Dragicevic, Chatchavan Wacharamanotham, Moving Transparent Statistics Forward at CHI, In: Proceedings of the 2017 CHI Conference Extended Abstracts on Human Factors in Computing Systems, New York, NY, USA, 2017-05-06. (Conference or Workshop Paper)
Transparent statistics is a philosophy of statistical reporting whose purpose is scientific advancement rather than persuasion. We ran a SIG at CHI 2016 to discuss problems and limitations in statistical practices in HCI and options for moving the field towards clearer and more reliable ways of writing about experiments, and received an overwhelming response. This SIG resulted in rough drafts of reviewer guidelines, resources for authors, and other suggestions for advancing a vision of transparent statistics within the field; this year, we propose a concentrated one-day writing workshop to develop those documents into a polished state with input from a diverse cross-section of the CHI community. |
|
Jaan Spitz, LocusStat: Augmenting charts with information influencing uncertainty, University of Zurich, Faculty of Business, Economics and Informatics, 2016. (Bachelor's Thesis)
Academic readers have to read scientific papers on a regular basis. To do this in an effective manner they tend to read the charts first. Unfortunately a lot of the visualized values and additional data that are needed to understand the graph are spread throughout the whole text, and are not readily accessible without effort.
To better interpret the chart, the reader has to search in the text for more information. This shift of locus of attention prohibits the reader to understand the chart in a timely manner, and increases the cognitive load unnecessary. This is annoying and consequently can lead to disregarding the text completely, which in turn makes the reader potentially severely misinterpreting the presented data.
This thesis shows that the LocusStat software can counter the above mentioned problems by augmenting the chats with data from the text. By adding interactivity to the graph to assist the reader in interpreting the visualized date. And by presenting the reader with links from the chart back to the highlighted text where the data is located.
|
|
Poornigha Santhana Kumar, Understanding How Novices Read Statistics, University of Zurich, Faculty of Business, Economics and Informatics, 2016. (Master's Thesis)
Research papers being an important source of information in all fields, it is spotted with statistical errors frequently. Statistics used in research papers is prone to errors in both author’s and reader’s side. To prevent statistical errors, both authors and readers behaviors were explored via literature survey. Since the results of the literature portrayed readers were relatively unexplored, readers’ side was studied in depth by conducting interviews. We revealed three major behaviors patterns exhibited by the readers while reading research papers. We focused on an obvious behavior exhibited by all readers: readers’ skips parts of research paper based on their interest. A prototype to aid the behavior was developed and evaluated. The evaluation results depicted that tools developed based behavior pattern was successful. Thus, this research proves designing tools based on behavior patterns helps readers in understand research papers and reduces statistical interpretation errors. |
|
Imanol Studer, Paper Tracking: A real-time algorithm, University of Zurich, Faculty of Business, Economics and Informatics, 2016. (Master's Thesis)
We want to change the way in which users interact with statistics in printed scholarly articles. We want to give them the opportunity to choose the representation of the data given in statistical charts. In this thesis we do the first step and present a system for extracting statistics from a printed document by detecting and tracking it using a head mounted camera. We show that it is possible to detect printed paper by looking for image regions with a high corner density. Furthermore, we match the paper against a database using SURF features to identify it. To extract the statistics, we search for figures and classify them using visual words based feature vectors. Paper recognition achieved an accuracy of above 90% with a database of 100 papers. Statistical graphics were found with a recall of 90% and a precision of 60%.
|
|
Nur Al-huda Hamdan, Florian Heller, Chatchavan Wacharamanotham, Jan Thar, Jan Borchers, Grabrics: a foldable two-dimensional textile input controller, In: CHI '16 EA: Extended Abstracts of 2016 ACM SIGCHI Conference on Human Factors in Computing, ACM, New York, NY, USA, 2016-05-07. (Conference or Workshop Paper published in Proceedings)
Textile interfaces can be ubiquitously integrated into the fabrics that already surround us. So far, however, existing interfaces transfer concepts, such as buttons and sliders,to the textile domain without leveraging the affordances and qualities of fabric. This paper presents Grabrics, a two-dimensional textile sensor that is manipulated by grabbing a fold and moving it between your fingers. Grabrics textile nature allows it to be integrated invisibly into everyday clothing or into textile objects, like a living room sofa. We describe the construction and the fold-based interaction technique of Grabrics sensor. A preliminary user study shows that Grabrics can be folded and manipulated from any arbitrary position, and it can detect 2D stroke gestures reliably. |
|
Philipp Wacker, Chatchavan Wacharamanotham, Daniel Spelmezan, Jan Thar, David A Sánchez, René Bohne, Jan Borchers, VibroVision: An on-body tactile image guide for the blind, In: CHI '16 EA: Extended Abstracts of 2016 ACM SIGCHI Conference on Human Factors in Computing, ACM, New York, NY, USA, 2016-05-07. (Conference or Workshop Paper published in Proceedings)
Today, persons with a visual impairment use a cane to explore their surroundings and sense objects in their vicinity. While electronic aids have been proposed to aid them, they communicate limited information or require a fixed position. We propose VibroVision, a vest that projects information about the area in front of the wearer onto her abdomen in the form of a two-dimensional tactile image rendered by an array of vibration motors. This vest enables the user to sense features such as shape, position, and distance of objects in front of her. |
|