Data Sciences

We often speak about data science, but recognize that our approaches to data vary widely depending on context, discipline, background, goals, and available tools. This page catalogues some of my interests and resources that I have found useful. My background is pure math to stats to data science, learning how to code informally, along the way. Early on–decades ago, even while an undergraduate–I realized I enjoyed combining math, biology and computers, and that I liked talking with people about their projects. What began as consulting evolved into collaboration, initially with colleagues in ecology, then agriculture, then animal models for human health. Lately, I have been interested in how communities learn to use data to understand their environment and health to make decision. All of these varied domains and interests have needs for data sciences insights. This page is an effort to pull together some practical tools in the context of my broader project to Document Digital Tools.

Examining relationships in complex systems presents many challenges. Many things are correlated, but not necessarily causally related. Measures over time and space are typically (auto)correlated, or correlated with each other. Sytems studies involve examining many different types of measurements, which raise issues of scaling, admixture or groupings, and even the conundrum that we don’t know what is useful to measure, or whether what has been measured properly reflects underlying processes. Some of these ideas are examined (albeit incompletely) in Making Sense of Data and Significance and Importance.

There is a tension between data sovereignty, the inherent rights of peoples to control access and use of their own data, and the open data movement. Many communities are exploring ways to use open, notably public, data to navigate local issues as well as more global challenges that impinge on their communities. I feel honored to be engaged in some of these conversations.

What is Data Science?

Making Sense of Data

My background is statistics, where data visualization and rigorous thinking about key questions and design are paramount. Other pages in this body address the mechanics of data visualization.

A question came up of how to relate rain and drought (or any other measures over time and space). I figured out how to create the plot I suggested from cumulative rain and drought. I used an example of two counties in SD, which are homes to Oglala Lakota (Pine Ridge in Oglala Lakota County) and Sicangu (Rosebud in Todd County). The published report SD Rain & Drought Analysis uses plotnine in python.

I think it is much better to show data explicitly (and do it early in a report) rather than talk about the concept. I also do not think correlation works well when considering processes over time and space, where there is substantial auto-correlation built in and variability due to the scale of measures. Cumulative measures are much better behaved for a variety of reasons.

Data Sovereignty

Open Data & Open Science

Table of contents

This site uses Just the Docs, a documentation theme for Jekyll.