Sunday, December 22, 2024

Welcome to Data for Public Health. I believe that anyone can learn and apply data science techniques to understand public health data, regardless of academic credentials. Cultivating the necessary tools and skills for such analyses are readily available within everyone's grasp. This site features example applications of data science, machine learning, and statistics using open-source software and public health data from sources around the globe.

Eventually, I’ll make an interactive intro video with a greenscreen and link it HERE.

The main purpose of this website is to feature my explorations of public health data from U.S. and global sources using open source platforms, but I also provide a series of brief articles spanning basic data science topics. These are intended primarily for my visiting students, as I currently teach health informatics at National University. Perhaps the curious will find some value in these offerings as well.

In addition to tags on each post, this website's content is laid out across five primary categories:
  • Intros: Broad walkthroughs of overarching data science concepts, such as public data, open source coding, and how learn data science.
  • Definitions: Brief clarifications on fundamental data science concepts.
  • Techniques: Summaries of various methods in statistics and artificial intelligence whereby data variability can be assessed, predicted, and understood.
  • Philosophy: Axiomatic presuppositions and intellectual contexts inherent to empirical data science.
  • Coding Projects: Explorations of our world through public health data leveraging statistical analyses, machine learning, and visualization libraries. 

To assist me with this website project, I developed a multi-agent system comprising seven chained LLMs under the direction of a supervisor AI. This architecture leverages several advanced technologies, including OpenAI, several Ollama models, Flowise, Grok, HuggingChat, Firecrawl, and LangChain, along with a Serper API web scraper and a couple additional, uncensored LLMs. (I find it easier to work with LLMs that haven't been programmed to fine tune me.) This also helped unlock translation of posts into other languages. If you're interested in learning more about this project, I can be reached through LinkedIn or via direct message on X.