Sunny Avry

Researcher & Data Scientist | Enhancing Education through Research and Data-Driven Insights | Specialist in Educational Advanced Analytics  research gate profilegscholar profilelinkedin profilegithub profile
Centre LEARN - Ecole Polytechnique Fédérale de Lausanne
PhD in Psychology - University of Geneva
Former teaching assistant in Computer Science and Instructional Technology
Former neuropsychologist

About me:

I am deeply passionate about applying research and data analytics to enhance education. Specializing in educational dashboards, I provide insightful visualizations that shape effective strategies. Beyond my profession, I explore the realms of web development and startup culture, constantly seeking innovative ways to leverage technology for practical solutions. My goal is to use my diverse skills to make a meaningful impact on education. Let's connect to discuss the intersection of psychology, data science, and education.


Monitoring training programs in digital competencies

As digital education curriculums continue to gain momentum globally, there's a growing necessity to enhance the digital skills of educators through tailored training programs. To ensure effective monitoring of these programs, the "EduNum" project, a digital education reform undertaken in the Canton of Vaud, has refined an existing model to oversee the entire process - from the training phase to the analysis of student learning outcomes. This model introduces crucial indicators for evaluation within digital education programs and fosters the utilization of an analytical methodology to augment the success of such ventures. The subsequent project aims to further validate the advantages of implementing a methodology derived from this model in the context of enhancing teachers' digital competencies and boosting students' learning outcomes. As a part of the 'EduNum' reform's initiatives, a critical evaluation was carried out to overcome the challenges that trainers encounter when they aim to appraise their training programs. This analysis gave rise to the conception of a web toolbox enhanced by AI language models named The Digital Training Companion, specifically designed to assist trainers in tracking their digital education training. For example, The Digital Training Companion provides an easy-to-use tool including key functionalities such as:

  • My Monitorings: This section allows you to create monitoring sessions including assessments of the different levels (e.g., immediate reactions, behavioral changes, etc.)
  • My Results: This functionality allows users to view and track over time all the results of their assessments.
  • My logbook: This functionality allows users to keep a logbook of all the modifications made to their training action in order to ensure its continuous improvement.
Download the manuscript

The Digital Training Companion is currently under development, so stay tuned for exciting updates and new information coming soon!

Collaborative problem-solving

Building on current theoretical frameworks and existing empirical findings, my doctoral research developed our understanding of the form and functioning of collaborative problem-solving (CPS). This programme of research led to two significant contributions evidenced in theoretical development and observable outputs.

(1) My research integrated theoretical approaches and findings from different domains of research related to the identification (form) of individual and group elements of goal directed collaborative interactions, that spanned cognitive (i.e., to look at communications focused on problem-solving) (e.g., Decuyper et al., 2010), motivational (i.e., considering the persistence of behavior to solve a problem) (e.g., Järvelä et al., 2008) and relational (i.e., reflecting the development of group relationships) (e.g., Hale et al., 2005) dimensions. Based on this integration of theory and research, I developed a new collaborative problem-solving (CPS) model to capture how these three dimensions interact with each other (functioning). This model (in revision) represents a dynamic process that illustrates how individuals engaged in CPS build and update mental models (Dillenbourg et al., 2016), i.e., individual in-memory representations, combining general knowledge and strategies (Miele & Scholer, 2018; Veenman et al., 2004) and incoming information from both self (via individual processing) and other group members (through observable outputs) regarding cognitive, motivational, and relational aspects of the CPS task.

(2) The second contribution of my doctoral research was to highlight an additional role of emotion processes in the functioning of CPS. Previous research has found that emotion modulates the personal and interpersonal processes in CPS. My research extends existing conceptualisations of the role of emotions in CPS suggesting that it represents a distinctive and pervasive phenomenon that cannot be easily assimilated to the socio-relational dimension of collaboration. My research demonstrated how emotions are clues that group members use to make real-time inferences regarding themselves and their collaboration partners. For example, confusion can indicate a difficulty in understanding an aspect of the problem, boredom can indicate low commitment, contempt can indicate little liking for the collaboration partner, etc. Group members can therefore enrich their cognitive, motivational, and relational collaborative models through the understanding of their own and others’ emotions and use them to regulate collaboration.

My thesis presents several experimental studies that provide empirical evidence of the impact of different emotional processes (i.e., subjective feeling, sharing of emotions, emotion regulation dispositions on the three core dimensions of CPS. This work has significant implications for training CPS to students and employers and that focuses on emotional awareness – to teach individuals to effectively regulate their own emotion and read others’ emotions during a collaborative task. This important and highly novel work has been published in peer-reviewed journals and proceedings of international conferences (e.g., Avry et al., 2020a; Avry et al., 2020b).


Beyond the Dichotomy between the Socio-cognitive and Socio-emotional Spaces : The Pervasive Role of Emotions in Collaborative Problem-Solving

Download the manuscript


The field of collaborative problem-solving has been gaining interest over the last decades. However, we are still far from getting a complete picture of its functioning. One of the reasons is undoubtedly its underlying complexity. Indeed, a compre- hensive understanding of collaborative problem-solving requires paying attention to various phenomena that dynamically interact when people try to solve problems together. The present thesis aimed at deepening the understanding of collaborative problem- solving at four main levels. The first contribution is an extensive review of the current state of research on various personal and interpersonal processes playing a role in collaborative problem-solving. To this end, we reviewed scientific contri- butions from different fields of research concerning the cognitive, motivational and relational aspects of collaboration. The second contribution is the construction of an integrative model that considers how these afore-mentioned dimensions inter- act during collaborative problem-solving at the personal and interpersonal levels. Moreover, the pervasive role of emotions as a source of information and regulation in each of these dimensions is also highlighted, challenging the classic dichotomy between socio-cognitive and socio-emotional spaces of collaboration classically pre- sented in the literature. All in all, this model is intended to provide a theoretical framework for further research in this domain. The third contribution concerns the study of some ways in which emotional processes influence collaborative problem- solving. Four studies explored the impact of self-experienced emotions, explicit sharing of emotions and emotion regulation dispositions on collaborative exchanges and the perception of different aspects of the collaboration. Finally, as a fourth con- tribution, we build on the findings uncovered in this thesis and the literature to propose new promising avenues for future research in this domain.


Teaching Assistant

TECFA - University of Geneva

Course: Digital learning and Distance Education (Master) with Prof. ass. Gaëlle Molinari

Examples of Instructional Software Prototyping (Supervision as part of the course ADID - Master MALTT)

2018 - Present
Logo UniGE

Teaching Assistant

Department of Computer Sciences - University of Geneva

Course: Software Engineering (Bachelor) with Prof. Philippe Dugerdil

2019 - 2020
Logo UniGE

Teaching Assistant

Department of Computer Sciences - University of Geneva

Course: Data Structure (Bachelor) with Prof. Stéphane Marchand-Maillet

2016 - 2020
Logo UniGE

Teaching Assistant

Department of Computer Sciences - University of Geneva

Course: Algorithms Programming (Bachelor) with Prof. Thierry Pun

2014 - 2018
Logo UniGE

Teaching Assistant

Distance Learning University Switzerland

Course: Methodology in Experimental Psychology (Bachelor) with Prof. ass. Gaëlle Molinari

2014 - 2018
Logo UniGE


PhD in Psychology - 2020

TECFA - University of Geneva

Logo TECFA   Logo UniGE

Master degree in neuropsychology - 2013

University Paris Descartes

Logo Paris Descartes

Master degree in cognitive science (Cogmaster) - 2012

Ecole Normale Supérieure Ulm

Logo ENS

Bachelor in Psychology - 2010

University Rennes 2 & University of Montreal

Logo Rennes 2   Logo Montréal

Contact Information

Sunny Avry

email: sunny.avry[at]
Phone: (+41) / (+41) 21.693.84.27

Collab Toolbox : useful tutorials for collaboration analysis (and more... :-))

Statistics for dyadic and multilevel analyses

Compute intraclass correlation to check non-independance of dyadic data (R)

Restructure a dyadic data table (Python)

Multilevel analysis (R) Download .Rmd file

Other statistics

Paired Samples Wilcoxon Test in R (R)

Compute a permutation test (R)

Mutiple comparisons correction with Bonferroni and Benjamini-Hochberg procedures (R)

Regression with a Single Ordinal Explanatory Variable and one Continuous Dependent Variable (R) Download .Rmd file

Regression with a Single Categorical Explanatory Variable and one Continuous Dependent Variable (R) Download .Rmd file

Regression with More than One Explanatory Variable (Multiple Regression) and one Continuous Dependent Variable (R) Download .Rmd file


Blur part or all of a video (ffmpeg)

Extract audio from video (ffmpeg)

Couper une vidéo en plusieurs parties avec FFMPEG (ffmpeg) (in french)

Text to speech

Convert text into speech with the Google Text To Speech API (Python)


Intégrer un notebook Jupyter dans un wiki (in french)

Coding Tips

Create a variable from a string and vice versa (R)

Manage environments and variables (R)

Regression assumptions checking (R)

Continuous training


Preparation for IELTS

The City of Liverpool College


Professional English writing B2-C1



Certificate of Advanced English C1



Research and teaching

Presentation training

SEA - Soutien à l'enseignement et à l'apprentissage - Université de Genève

Logo sea

Research communication

SEA - Soutien à l'enseignement et à l'apprentissage - Université de Genève

Logo sea

Enhancing Ph.D. skills

SEA - Soutien à l'enseignement et à l'apprentissage - Université de Genève

Logo sea


Multilevel modelling online course

University of Bristol

Logo uniBristol

The Data Scientist’s Toolbox

Coursera - Johns Hopkins University

Logo coursera   Logo jhu

Computer Sciences

R programming

Coursera - Johns Hopkins University

Logo coursera   Logo jhu

Data Management

Research Data Management and Sharing

Coursera - University of North Carolina at Chapel Hill & The University of Edinburgh

Logo coursera   Logo university_north_carolina   Logo university_edimburgh

Other stuff

My Memrise courses : Fun, fast ways for English Speakers to effectively learn and memorise vocabulary.