---

Introduction

I am not a self-promotional kind of guy, but I have to tell you that I am a man on a mission. When I began this mission five years ago, I had no idea that it would consume me like it has. Some people see me and are envious of my commitment, while others sympathize with me and think I should be committed!

I had no idea that my understanding of global warming would increase to the point that I may now understand global warming better than many professionals that are paid to study the topic.

The reason I can state this is that I have now unleashed the full power of Alteryx and Tableau on the data, and I let the data tell me stories through a thorough examination of temperature changes over the past 60 years or so.

As shown in Figure 1, I want to understand the detailed changes across space and time that have created the temperature anomaly curve shown between 1960 and 2017.

With a global temperature increase of 1.69 degrees F, I have been trying to understand how big the daily variations of temperature are at all points across the earth.

Even though daily weather fluctuations are hard to predict, what I have discovered is that buried in the data are mathematically significant time series temperature trends that beautifully describe what is happening across our planet.

If you want to learn the truth about global warming, get ready to strap it on (as my friend Norm used to like to say!). Come along for a ride with me to learn about how warming and cooling patterns have developed over the past six decades, and how these patterns change throughout each day of the year.

Interrogating the Data To Detect Global Warming Signals

The insights I am about to share took me a lot of time to uncover. The insights occurred as I poked and prodded the data over the past few years in different ways. I used different levels of aggregation, different spatial groupings, and I did a gazillion computations (or rather, Alteryx did that work!).

I didn’t know what I was going to find, but once I started seeing the patterns, my fire was lit. Now that fire continues to burn and gives me energy and enthusiasm for continuing this quest.

When I thought about temperature changes over time from a new perspective (i.e., the daily perspective) I made some interesting discoveries. I could see in more detail what has been happening to our temperatures compared to the work I did at the monthly level of analysis. The monthly analysis showed me the basic patterns, the daily perspective gave me the details.

The daily perspective isn’t an obvious choice to use for this type of analysis because of the variations we know exist in the weather. It turns out, however, that the global warming signals are easily detected when you use a daily perspective as measured and computed across decades.

A daily perspective is defined as follows. You create a time series of temperature data for any day of the year. You pick a starting day, such as Jan 1, 1960, and you pick a temperature measurement such as maximum daily temperature (Tmax). Your next data point comes 1 year later (Jan 1, 1961) after the earth has revolved around the sun and returned to the same position. By building up a nearly six-decade record of daily temperature readings, I could see what has happened over time and space. Now do that for every day of the year for every monitoring station on earth that has enough data, and the global picture of what is going on becomes much clearer.

By computing daily temperature trends across time, measured year by year (e.g., Jan 1, 1960, Jan 1, 1961, Jan 1, 1962, etc), I saw the emergence of the temporal global warming signals. By plotting the data on a global map, I could see the spatial trends over time. These two activities have allowed me to uncover what I have been searching for.

---

When the data is structured like this and millions of temperature trend models are computed using about ninety million temperature readings, the stories become clear. The visualizations of this data explain the deleterious environmental effects we are seeing across this beautiful planet. When proper visualizations of the data are prepared, the spatial and temporal temperature changes created by global warming come to life.

In one sense, I have completed a massive data mining operation that has produced some serious gold.  If you would have asked me five years ago what I expected to find out by doing this work, I would have been completely wrong! What I have learned has transformed me. I hope I can help you learn a little about what is happening on this beautiful planet of ours, too.

Processing the Data

I dare say that very few people have had the wherewithal to attempt to process the amount of data that I have, in the ways that I have. I have been incessant in my pursuit of improving my understanding and comprehension of global warming. I have gone bonkers. I still refuse to quit, despite feeling like I can now explain what has happened and what will likely continue to happen over time.

If you want to see how this work has been done, subscribe to this blog via email because I will be publishing more information on how the data gets processed in Alteryx and visualized in Tableau.

This Data is From Planet Earth

Now it is time for me to share some data with you. I have not held anything back. I am offering to you the full complement of the data, with no apologies for kicking your ass by giving you a serious data overload.

If you want to see what is happening throughout space and time due to global warming, you now have no excuses. It is incumbent upon you to learn by studying the daily dashboards and seeing for yourself what I have learned.

Let me introduce you about 180 million temperature readings (Tmax, Tmin) from a planet called Earth. I’ll throw in another 90 million daily temperature ranges (Tmax-Tmin), just for fun. These temperatures have been organized by day of the year.

Mathematical trend models have been used to calculate the total amount of warming or cooling that has happened at each of the monitoring stations around the planet that have data spanning from 1960 to 2018. These trend models are based on real temperature data recorded at monitoring stations, not simulated data. This is the true story of global warming.

To understand the full power and beauty of these Tableau dashboards, I offer the following video. Yes, at 12 minutes long, it is more than I wanted it to be. However, I think I deserve a little extra time considering the effort I have made in creating this beast.

---

How to Use the Daily Digest of Worldwide Global Warming Dashboards

I know it is outlandish to do this, but it is time that we cut through the BS and have intelligent discussions of what global warming really means. With 365 individual Tableau Public dashboards, I think Tableau is going to send me a bill for storing and sharing this story!

---

Describing The Incredible Tableau Dashboards

There will be a dashboard for each day of the year, once my work is complete. For each day, there are approximately 250,000 daily maximum temperatures (Tmax), daily minimum temperatures (Tmin) and the daily temperature range (Trange). There are up to 5,200 monitoring stations included in each of these data sets. You will be able to investigate the spatial and temporal changes in Tmax, Tmin, and Trange for any location on earth that interests you, for any day of the year.

As of the publication date, I have not completed uploading all of the dashboards to Tableau Public. I have only uploaded the dashboards for the months of July and August. It is going to take me some time to upload the rest of them. The reason I started in July is that this is where we are as I write this on July 6, 2019. To show how these can be used, I offer the following four visualizations.

Figure 2 is the interactive dashboard for today – July 6. The overall temp change is 2.1 deg F, which is higher than the 1.69 global average. This number is found in the lower right panel, in the yellow highlighted block. You can scroll up and down this list to see how many stations are included in the dashboard, via the Index field. In this case, there are 4,190 stations that have between 40 and 59 years of data on July 6, for the timeframe between 1960 and 2018.

---

It looks like the Pacific Northwest and California (Figure 3) has seen a 3-degree change, while the northeast US states (Figure 4) has should expect warmer conditions than 1960. The midwest US states show cooler temperatures.

---

---

As shown in Figure 5, Europe is much hotter than it was in 1960. Many stations in France, Germany, Spain, Italy, and other countries are showing more than 10 degrees of additional heating compared to 1960. The average for this region is 5.6 deg F, which is similar to the upper east coast of the US.

---

Easy Access to the Daily Digest Dashboards of Global Warming

If you see colored hyperlinks in the following two tables, you can click on the links to launch the daily digest of worldwide global warming dashboards for those dates. You can also download them from Tableau Public for your own usage. So far, I have only uploaded July-August data. Be sure to watch the video earlier in this article to see how much functionality exists in these dashboards.

The Daily Dashboard For The Second Half of The Year

Day	Jul	Aug	Sep	Oct	Nov	Dec
1	Jul-1	Aug-1	Sep-1	Oct-1	Nov-1	Dec-1
2	Jul-2	Aug-2	Sep-2	Oct-2	Nov-2	Dec-2
3	Jul-3	Aug-3	Sep-3	Oct-3	Nov-3	Dec-3
4	Jul-4	Aug-4	Sep-4	Oct-4	Nov-4	Dec-4
5	Jul-5	Aug-5	Sep-5	Oct-5	Nov-5	Dec-5
6	Jul-6	Aug-6	Sep-6	Oct-6	Nov-6	Dec-6
7	Jul-7	Aug-7	Sep-7	Oct-7	Nov-7	Dec-7
8	Jul-8	Aug-8	Sep-8	Oct-8	Nov-8	Dec-8
9	Jul-9	Aug-9	Sep-9	Oct-9	Nov-9	Dec-9
10	Jul-10	Aug-10	Sep-10	Oct-10	Nov-10	Dec-10
11	Jul-11	Aug-11	Sep-11	Oct-11	Nov-11	Dec-11
12	Jul-12	Aug-12	Sep-12	Oct-12	Nov-12	Dec-12
13	Jul-13	Aug-13	Sep-13	Oct-13	Nov-13	Dec-13
14	Jul-14	Aug-14	Sep-14	Oct-14	Nov-14	Dec-14
15	Jul-15	Aug-15	Sep-15	Oct-15	Nov-15	Dec-15
16	Jul-16	Aug-16	Sep-16	Oct-16	Nov-16	Dec-16
17	Jul-17	Aug-17	Sep-17	Oct-17	Nov-17	Dec-17
18	Jul-18	Aug-18	Sep-18	Oct-18	Nov-18	Dec-18
19	Jul-19	Aug-19	Sep-19	Oct-19	Nov-19	Dec-19
20	Jul-20	Aug-20	Sep-20	Oct-20	Nov-20	Dec-20
21	Jul-21	Aug-21	Sep-21	Oct-21	Nov-21	Dec-21
22	Jul-22	Aug-22	Sep-22	Oct-22	Nov-22	Dec-22
23	Jul-23	Aug-23	Sep-23	Oct-23	Nov-23	Dec-23
24	Jul-24	Aug-24	Sep-24	Oct-24	Nov-24	Dec-24
25	Jul-25	Aug-25	Sep-25	Oct-25	Nov-25	Dec-25
26	Jul-26	Aug-26	Sep-26	Oct-26	Nov-26	Dec-26
27	Jul-27	Aug-27	Sep-27	Oct-27	Nov-27	Dec-27
28	Jul-28	Aug-28	Sep-28	Oct-28	Nov-28	Dec-28
29	Jul-29	Aug-29	Sep-29	Oct-29	Nov-29	Dec-29
30	Jul-30	Aug-30	Sep-30	Oct-30	Nov-30	Dec-30
31	Jul-31	Aug-31		Oct-31		Dec-31

The Computation of the Trend Models

I promised a long time ago to not write any more about mathematical modeling Tableau! I said that because I went deep on this topic years ago: (math modeling in Tableau). I have to say that my previous statement is now going to be a lie because Tableau is still developing great software with awesome capabilities.

For the first time (Tableau version 10.2 or newer is required), I have been able to do what I wanted to do for the past six years! I can easily do real-time, dynamic temperature trend modeling for thousands of monitoring stations. I can pick subsets of stations and get insights faster than I can write these words.

Here are the basics of developing trend models for time series data for a variety of locations (i.e., monitoring stations, stores, etc). I used fixed LOD calculations to do the work in a five-step procedure. This is so easy now that anyone can do it!  Thanks, Tableau – you guys rock!

---

Step 1: Pick a Temperature Variable via a Parameter (Tmax, Tmin, Trange) 

The calculated field is called: [Temp_Variable]

if([Temperature Variable]=”Tmax”) then [TMax (deg F)]
elseif([Temperature Variable]=”Tmin”) then [TMin (deg F)]
else [TRange (deg F)]
end

Note: The variables Tmax, Tmin, and TRange are measures in my data set.

---

Step 2: Find the Minimum Date of Time Series Data by Station

The calculated field is called: [Min Date LOD]

{fixed [Station ID]: min([Date])}

Note: The variables Station ID and Date are dimensions in my data set. Station ID is the code for the temperature monitoring stations across the earth.

---

Step 3: Find the Maximum Date of Time Series Data by Station

The calculated field is called: [Max Date LOD]

{fixed [Station ID] : max([Date])}

---

Step 4: Compute the Slope of the Trend Across Time by Station

The calculated field is called: [Slope LOD]

{fixed [Station ID] : covar(INT(DATETRUNC(‘year’,[Date])),([Temp_Variable]))}

/ {fixed [Station ID] : var(INT(DATETRUNC(‘year’,[Date])))

Note: I think Tableau 10.2 or newer is required to access the covariance and variance functions.

---

Step 5: Compute the Total Temperature Change Across the Time Frame

The calculated field is called: [Temp Change LOD (F)]

[Slope LOD]*([Max Date LOD]-[Min Date LOD])

---

You can download any of the daily dashboards to retrieve this formulation.

Insights to Changes in the United States

In a previous article, I described how I aggregated this data by state to develop an understanding of the changes going on in the US. The following charts are a summary of that work. That work was built using the data presented in this article.

This slideshow requires JavaScript.

---

Previous Work

I have done a lot of work on this topic. Click here to review my publications on global warming and to learn why I have been doing this work. If you want to know about the source of this data, go to the Phase 1 work in the citation given above.

Thanks for reading.