https://www.gravatar.com/avatar/6f857b907c2edf7d71d705e6de716006?s=240&d=mp

Roberto Aguilar

Analyst @ McKinsey & Company, AI & Gaming

Welcome to my portfolio website, any views, material or statements
expressed are mines and not those of my employer

GenAI-Driven Pokémon Go Companion App: Augmenting onboarding experience with RAG

Abstract: This article introduces an E2E application aimed at enhancing the onboarding experience for new Pokémon Go players through a Langchain-framework-based app that integrates with the OpenAI API, leveraging GPT-4’s capabilities. Central to the application is a GenAI-powered Pokémon index, designed as a companion to reduce new user friction and deepen understanding of the game’s universe. By employing a Retrieval-Augmented Generation system with FAISS, Meta’s vector database, the app ensures information retrieval.

Optimizing User Retention in Brawl Stars: Unsupervised Machine Learning Approach to Cohorts Retention Metrics

Abstract: This study presents a comprehensive investigation into the development of a player retention pipeline and the establishment of user cohorts for the game Brawl Stars through the application of unsupervised machine learning models. The primary aim is to enhance the analysis of user retention by implementing advanced cohort classification techniques and the generation of retention metrics. To achieve this, the research utilizes data obtained from the Brawl Stars public API, applies KMeans clustering for the categorization of players, and employs PySpark for data processing and analysis, with an emphasis on the use of Kedro for the efficient assembly of the data processing pipeline.

Assessing the Impact of Patch Releases on Digital Assets in DOTA 2

Abstract: This study examines the effect of a game patch on hero balance in Multiplayer Online Battle Arena (MOBA) games, focusing on changes in hero damage output. Utilizing ANOVA tests and statistical analyses, including Kruskal–Wallis ANOVA, Mann-Whitney U test, Kolmogorov-Smirnov test, and Bootstrapping in Python, data from ranked matchmaking sessions were analyzed to assess the statistical significance of damage output changes according to the official changelog. Findings reveal Shadow Fiend as the standout hero with increased gold collection, tower damage, and victories due to the ‘Requiem of Souls’ ability’s enhancement with the “Aghanim’s Scepter” in patch 6.

Exploring the Impact of Control Mechanics on User Experience in Fallout New Vegas

Abstract: This publication shows a comprehensive statistical examination of control mechanics within Action RPGs, exploring the hypothesis that alterations in control mechanics substantially affect gameplay outcomes, specifically in terms of quest completion and enemy kills, across varying player experience levels. Utilizing data extracted from the VPAL mod, the research employs a methodological approach that includes data preprocessing, and the application of inferential analysis via t-tests and Mann-Whitney U tests, executed within a Python environment.

Level Balancing: Game Difficulty data-driven adjustments in Candy Crush

Abstract: This study explores the impact of game difficulty on player engagement within mobile Puzzle Games, positing that the level of challenge significantly influences player retention and satisfaction. Utilizing Python and Plotly for data preprocessing and visualization, the study analyzes player attempts and success rates across various levels, based on a dataset provided by Rasmus Baath. The findings reveal a generally well-balanced difficulty spectrum, advocating for the implementation of dynamic difficulty adjustments to optimize engagement.

A/B Testing Study on the Impact of Time Gate Positioning in Cookie Cats

Abstract: This study explores the influence of adjusting the positioning of a “time gate” within the mobile game Cookie Cats, transitioning it from level 30 to level 40, on player retention rates. Predicated on the hypothesis that repositioning the gate would significantly impact user retention, the study gathered and analyzed data encompassing user IDs, game version details, rounds played, and retention metrics on the first and seventh days post-engagement. Employing exploratory data analysis alongside robust statistical methodologies, such as bootstrapping, to evaluate the collected data, the findings reveal a disparity in the retention rates on the first day, with a slight preference towards the control group, wherein the gate remained at level 30.

Video Games History explained with Pandas

The industry of video games revenues is reaching the $173.7 billion in value, with around 2.5 billion users enjoying them worldwide, with a forecasted value of $314.40 billion by 2026 according to Mordor Intelligence.

Impressive facts, right? Nowadays this market is no longer considered a simple hobby for kids, it has become a constantly growing giant which attracts more and more customers as it takes advantage of the growth of streaming platforms. But this industry, as we well know, is taking more fields, born from ambitious expectations such as the Nintendo World Championships in the 90’s to what today many have adopted as a lifestyle also known as Esports.

In this exploratory analysis, we will also see one of the events that marked the industry in the 80s. Here is a preview.

Game Development Team Structure

In the next portal, you will find a frame with a diagram. This one displays the distinct roles and sub-roles in a video game development team. This section intends that as you read the Stakeholders section in each of my projects, you can learn about the operations, roles, and skills that each team member must possess. For this, you can read the details at the bottom of the card made for each.