Introduction
Bloggers like to count the articles they have produced. Article numbering gives us a way of measuring our progress towards becoming better storytellers, completing research, and helping educate others. We try to save special topics for significant article numbers, like 100, 200, or even 400! This article happens to be my 400th story in this blog. I believe that it contains a few inciteful topics, so let’s get started.
Do You Like to Read?
I like to read, and I enjoy reading a great book that takes many days or weeks to complete. Books by John McPhee or James Michener take time to complete because the stories typically span significant amounts of space and/or time. The research in both McPhee’s and Michener’s books is very detailed and generally took the authors (and Michener’s research teams) years to complete. I enjoy reading these books because the stories unfold in my brain as I visualize the stories’ landscapes, characters, and logical plots. Books like these are better than movies can ever be.
Last month, I read an excellent book called “Masterpiece” by Dean Stoecker. Much like the authors I mentioned above, Dean told a tale that spanned over two decades as he and his two co-founders (Libby Duane Adams and Ned Harding) built a company called Alteryx. Without the work of these three people and countless others, the global warming research I completed would not have been possible in the time I had to do it. The last decade of my career would also have been much less productive if I did not have Alteryx.
Eight years ago (in 2014), I decided to try to read an imaginary book. I titled this imaginary book: “Is Global Warming Real or Is It Fake?”. I could see the book in my mind, and I was eager to learn the answer to the question posed in the title.
When I first opened this very thick imaginary book, I was startled to find that the book contained no text! Instead, I encountered an unexpected type of language. I found myself face-to-face with millions of lines of numerically-encoded climate data. I quickly realized I needed to learn to read and decode this climate data to find the answer to my question.
Luckily for me, the book contained all of the information needed to answer my question. The book contained billions of numbers recorded from sites in many countries. This data had been collected and collated by hundreds of thousands of workers over the past two centuries. I felt privileged to have access to this information. I knew this would be a case that supported the ideas I wrote about in 2016.
I also felt a little perplexed because I knew that it would take a lot of work to find the answer to my question. The symbols used on Egyptian cuneiforms were not unlike the lines of data I had to decode. I imagined that Egyptian archeologists felt this way when they saw their first cuneiform (Figure 1) and wondered what all those symbols represented.
At the beginning of this work, I remember thinking that it might not be possible for me to explain or demonstrate global warming impacts systematically. I expected the data to show chaotic temperature changes over space and time. I believed that non-linear atmospheric characteristics would lead to really messy data that would be hard to explain, if not impossible. These negative expectations helped propel me through the years of doing this work because I was excited once I began making explainable discoveries.
The Most Important Data Decoders I Have Today
When I chose to start my global warming research in 2014, I had no idea how long it would take or how it would unfold over time. Looking back on my journey, however, I can see how the software tools I used allowed the data to tell an exciting story that has been full of surprises and insights that I never imagined would be possible to uncover.
Since I completed this work during my free time on nights and weekends, I had intrinsic limits on what I could accomplish. To reach my goal, I needed fast software that could allow me rapidly process and visualize data. My two tools are:
- I use Alteryx to quickly write computational code to produce custom data sources that could be analyzed and visualized;
- I use Tableau to rapidly visualize large amounts of data, with the flexibility to show temporal and spatial trends at multiple scales.
My ability to complete this work was totally dependent upon these tools. Thank you Alteryx and Tableau companies!
How The Research Occurred
When I think of how I completed this global warming research, one important insight emerges. That insight is this: I used data to uncover and describe many aspects of global warming that I had never seen before.
In 2014, I first had to understand the data structures and qualifiers used to store the information. Accomplishing this took me over 100 hours to learn, test, and document (see phase I results). Once I felt comfortable with my working knowledge of this data, I employed basic data mining techniques to uncover the stories hidden in the data during the first seven years of research. By 2021, new data structures required new data processing codes. Luckily, Alteryx is such an amazing platform for rapid prototyping and development that the new workflows were very easy for me to build.
Data mining is a process that allows you to explore data by asking questions about the data and then working with the data to find the answers to those questions. These questions can involve different scales for space and time. Generally, I moved from big space and time scales (i.e., monthly temperature changes across decades) to more detailed time and space scales (i.e., daily insights at the country, state, and monitoring stations).
Multiple research iterations were needed for me to understand how climate changes could be characterized and understood at different scales. To find answers to each new question I created, I built Alteryx workflows and specific Tableau dashboards to visualize the data. Over time, I built dozens of Alteryx workflows and Tableau dashboards. Each Alteryx workflow I built (Figure 2) performs specific tasks that allow me to generate the custom data sources that I have visualized in Tableau (see next section).
This iterative approach allowed me to comprehend climate changes in space and time across multiple scales. Answers to one question allowed me to formulate new questions. There was a natural flow to the research that allowed me to uncover interesting global warming insights. I wish I could say that I had a plan to do this work, but that would not be true. Curiosity, intuition, and an insatiable thirst for knowledge drove this work.
How Tableau Helped Me Understand and Uncover the Stories
At each stage of the research, I built new Alteryx workflows and Tableau dashboards to interrogate the worldwide climate data so that I could answer the questions I was creating. I have learned that building simple and targeted dashboards is essential when making fast progress towards achieving data comprehension. Spending a lot of time doing “pretty-printing” of dashboards is not a luxury I had while doing the research. It was more important for me to create quick data visualizations than create visual masterpieces. This is one of the primary benefits of Tableau – it works with you to get things done quickly. I like to say that I use Tableau exactly the way it was designed to be used.
The slideshow below illustrates Tableau figures and dashboards that I developed over time. The titles used in each example explain the purpose of the visual. One insight from this slideshow is that there is a lot of consistency in how I chose to display the data, even with the analysis spanning eight years and multiple topics being covered. You can click here for a pdf of this information.
What is the Significance of This Work?
When I began to decode the imaginary global warming book, I had a modest goal. I wanted to understand if global warming was real or fake. I wanted the data to tell me the truth, rather than listening to all the jabber that exists on the topic.
I can now proclaim that the answer to my question is that global warming is real. Climate change is also real. The significance of what I have discovered is multi-faceted, with several surprises included. Here are the highlights.
- There are statistically significant, worldwide heating and cooling patterns that have developed across the past 60 years. These patterns coincide in time with the significant rise in atmospheric CO2 levels.
- The temperature change patterns can be computed for each day of the year. Certain days have significant patterns that have developed, while other days are very quiescent without much temperature change detected. Click here to view a movie of the patterns that have developed in the United States.
- These patterns can be identified in many countries across the world. Click here to view a movie of the patterns that have developed in Europe. The patterns are best delineated where the climate monitoring network is well-established, with consistent data collection over the past four to six decades. One of the interesting differences between the US and European patterns is that the US has larger cooling zones across the central US, whereas the European patterns are more heat-related. This could be true because European weather is more correlated to ocean temperatures. With our oceans absorbing more than 90% of excess heat, it’s possible that European countries are beginning to feel the effects of the Atlantic ocean beginning to release some of the excess heat.
- Many of the temperature change patterns seem to have developed due to perturbations in the jet stream over time. In the future, I hope to understand this better.
- The largest temperature changes detected are in the northern latitudes of the northern hemisphere. Significant temperature changes are also happening near the south pole, although the monitoring network is smaller in space and time in the data sets I have used.
- The reason the largest temperature changes are occurring at the poles most likely has to do with the ratio between atmospheric CO2 levels and the total water vapor in the air. The poles have less atmospheric water vapor, which causes exaggerated heating due to the proportionately higher CO2 in the air.
- When significant temperature anomalies occur such as a deep freeze in Texas (Feb 2021) or extreme heating in Washington State (Jun 2021), there typically is a strong correlation between the daily temperature patterns identified. This has been true for both heating and cooling anomalies. These correlations lend legitimacy to the temperature change patterns. The magnitudes of extreme weather events are increasing over time, however.
- If the temperature patterns are to be believed, then it should be possible to make future temperature predictions using the models that initially detected the patterns. Although predicting future temperatures is typically inaccurate beyond a couple of weeks, the models built by this work have been shown to make accurate future temperatures more than 9 months into the future! Moreover, the accuracy of the predictions has been shown to be better than over 20 AI/ML time series prediction methods.
Future Work
I have already continued to perform research beyond what I have discussed in this article. What this means is that I will likely continue to uncover insights that are either poorly understood or currently not known. I would like to extend some of my predictive work and the geospatial visualizations I can produce to better elucidate this unfolding story.
Until next time, thanks for reading!
