Vid Stropnik.com

My personal Data Science Portfolio

temporal set prediction with gnns

Co-authored a comprehensive and engaging blog-tutorial (with included source code) on an advanced method of temporal set prediction using Graph Neural Networks as part of Stanford's CS224W course.

network analysis

With colleagues Vuk and Marusa, I studied the effect of Key Match Events on Football Passmaps. An article about our findings is available at Arxiv.

Time series analysis

With colleagues Tim and Luka, I implemented a fully serviceable time series analysis dashboard as part of the 2021 Data science Project competition winning product, Impress, in collaboration with Celtra.

report_VidStropnik.pdf

Exploratory Data Analysis

In the beginning of my studies, I performed a fun exploratory data analysis on whether advertising during the holidays pays off, based on Celtra's advertising data.

NLP: Summarization and Hypernym detection

IS2020_Volume_A - SLAIS-110-113.pdf

topical clustering method comparison

A short article at the slovene AI conference on the main contributions of my Bachelor's thesis (w/Zoran Bosnic and Evgeny Osipov)

VidStropnikSU.pdf

Topic labelling and Hypernym discovery

Some additional work in the related field, neatly packed into a quick overview of the domain, done as part of a course project.

Santorini_AI__Report.pdf

Planning &

Game AI

Developed a fully playable game of Santorini with an intelligent AI agent using the Minimax algorithm with Alpha Beta Pruning. Check it out and try it for yourself by pulling from our repository. (w/ colleague Luka)

OWN IMPLEMENTATIONS OF MACHINE LEARNING approaches

Implemented several key machine learning approaches, including trees (Bagging, Random Forests), logistic regression, kernelized ridge regression, support vector machines, and artificial neural networks from scratch. In the reports, enclosed in the ML-Showcase repository, the approaches are also tested and evaluated on real-world datasets. In said reports, I also provide insights into crucial data science concepts, like loss estimation and the quantification of uncertainty.

Connect with me!

I'd be really happy to get in touch with anyone who shares my interests and chat all things data! Feel free to write to me on LinkedIn or write to me at vid@stropnik.online.

Non-Data Science projects of note

Gamified mobile experience

Wore many hats (Content direction, Full Stack programming, Copywriting) in the development of the AR powered gamified heritage navigation app KulTura. (Available on Google Play, w/LTFE UL)

BACKEND JAVA PROGRAMMING

Worked on the Server side of a proof of concept application that Augments a football viewing experience with sports data in AR using Microsoft HoloLens. (w/ LTFE UL)

web accessibility

Coordinated a project on Web Accessibility, funded by the European Social Catalyst Fund. Evaluated existing WA compliance evaluation methodologies and organized a successful workshop.

event organization, pr, marketing

Organized 200+ events, often as the head of public relations or communication planning, with several local NGOs. Most notable include open-air concerts attended by thousands. Other roles include social media ad manager and project lead.

Editing & VOICE OVER

I've collaborated on over 20 videos in different parts of the production pipeline: Voice Over, Scripting, Direction, Storyboard art, Animation and Editing.

View some of the videos above.