Francis

Francis

Digital Transformer, Data Scientist, Systems Architect, Software Engineer | Specializing in Digital Transformation, Data Science, Data Analytics, Machine Learning, and System Design

Hi there! Welcome to my portfolio! Here are some samples of my work that I would love to share with you:

Comprehensive Medicare Data Analysis for Insurance Companies

Published:

In this study, the team leveraged Big Data techniques to analyze the CMS Medicare data and provide valuable insights for insurance companies to optimize their policies and coverage options. By effectively handling the large-scale dataset, we were able to uncover patterns in healthcare service utilization, payment dynamics, and drug claims among insured policyholders, leading to actionable recommendations for improved healthcare delivery.

Artificial K-Intelligence

Published:

Using Explainable AI, such as the Shap method, can provide valuable insights into the factors contributing to the popularity of K-dramas, which are renowned for their intricate and multifaceted storylines.

Pressure’s off the Menu! Enjoyable Plating for Healthy Living

Published:

Applying clustering and recommender system techniques, such as collaborative filtering, can enable the development of a personalized recipe recommendation system that takes into account both culinary preferences and nutritional requirements, particularly for individuals with hypertension.

A World of Flavors: Exploring Cities and Other Food Choices

Published:

By combining collaborative and content-based filtering approaches, it is possible to develop a restaurant recommendation system that suggests the most highly rated and similar restaurants in a given location. This system can take into account both the user’s past ratings and preferences (collaborative filtering), as well as the restaurant’s characteristics, such as cuisine, ambiance, and customer reviews (content-based filtering).

Uncovering Hidden Clusters: Analyzing the Bureau of Customs Import Dataset

Published:

Applying various clustering techniques, including representative, hierarchical, and density-based methods, can help identify potential clusters in the importation of goods by the Bureau of Customs. By uncovering patterns and similarities in the data, these techniques can provide insights into how to optimize the importation process, improve risk management strategies, and enhance revenue collection efforts.

Paws and Reflect: Which Dog Should I Select?

Published:

The combination of Web Scraping and Dimensionality Reduction techniques can facilitate the analysis and selection of dog breeds. By using an API for breed information, web scraping can collect a large amount of data on various dog breeds. Dimensionality reduction techniques, such as Principal Component Analysis (PCA), can then help identify the most important factors that differentiate the breeds. This can ultimately simplify the decision-making process for selecting a dog breed that best fits an individual’s preferences and lifestyle.

WanTED: Data Science

Published:

Using Web Scraping with the Youtube API and Text Vectorization techniques can help identify the top 10 data science videos on a particular Youtube channel. By web scraping through the Youtube API, a large amount of data on the videos can be collected, including titles, descriptions, and other metadata. Text vectorization can then be applied to transform the textual data into numerical vectors that capture the semantic meaning of the text. By analyzing the vectorized data, the top 10 data science videos on the channel can be identified, providing a valuable resource for individuals interested in the field.

Dota 2 Match Analysis: Unleashing the Potential of LightGBM Machine Learning Techniques

Published:

Machine learning models, such as LightGBM, can be applied to accurately predict the winner in competitive gaming. LightGBM is a high-performance framework that can quickly build models with high accuracy. By using historical data on gameplay, player performance, and other relevant factors, a LightGBM model can be trained to predict the outcome of a match. This can be valuable for gamers, esports enthusiasts, and betting companies looking to make informed decisions based on data-driven insights.

Magic the Gathering: Winning Not by What (Colors) You Play, but by How You Play

Published:

Data mining and wrangling techniques can be used to gain insights into the complexity of Magic: The Gathering and how it affects gameplay to increase wins. By collecting and cleaning data on various gameplay elements, such as card types, player strategies, and game outcomes, data mining and wrangling can help uncover patterns and relationships that contribute to winning strategies. This can provide valuable information for players and game designers looking to optimize gameplay and enhance the overall gaming experience.