This July (or this July too?) was hot. The heat of this time of year is nothing new, but these days when it's been so hot for a long time that it's a hassle to go outside.
NASA news according to, The average global temperature in July this year was the highest ever recorded in July. It is stated that July 2023 was 1.18°C higher than the average July temperature from 1951 to 1980.
Some people may think, "Is it just 1°C higher than it was a few decades ago? It's not a big deal," but a 1°C rise in temperature will cause more large typhoons, floods, and droughts. It is said that the possibility of a dangerous situation for human beings will increase.
Therefore, in this blog, I will try to visualize how the average temperature of the world is changing, mainly using heat maps.
Average global temperature over the past 30 years (based on 1951-1980)
On the NASA page, the temperature difference in the corresponding month is published as data based on the temperature from 1951 to 1980.
So, let's show the monthly temperature difference over the past 30 years (1994-2023) using a line graph and a smooth line. The line graph shows the minimum (=0.03) and maximum (=1.37) temperature difference during this period. also shows
Since the minimum value is 0.03, there are no months below the standard (1951-1980) in the past 30 years. If you have 30 years of data, it would be nice to have a month or two that fell below the standard, but this is surprising.
Looking at the smooth line (blue line), It is clear that the temperature difference is on the rise. However, if you look at it locally, you can see that the rise was large from 2014 to 2017, and that it has not risen that much in the past few years.
Create a heatmap and set a color gradient
Since it is difficult to compare by year or month with this line graph, heat map Let's create a . It's a "heat" map because it visualizes temperature.
Using JMP's Graph Builder, setting the year as the vertical axis, the month as the horizontal axis, and the temperature difference as the color, the following heat map is drawn with the default settings.
A legend is displayed on the right side of the graph, indicating that the darker red months have higher temperatures than the reference.
Conversely, it should be noted that it is not true that "the darker blue months indicate lower temperatures than the standard".
As shown earlier, in the past 30 years, all years and months have exceeded the standard temperature.
However, since blue has a cool image, some people may have an image that the darker the blue, the lower the temperature than the standard.
So right click on the legend and [Gradation] to change the color settings.
The main changes are the two points circled in red in the figure below.
- Change color theme : In this example, colors of similar colors, the larger the value, the It is a good idea to choose a theme with darker colors. (Here, the color theme of yellow → red is selected.)
- Change minimum and maximum values : In this example, we know that it can only take values greater than or equal to 0, so we set the minimum value to 0.
This is a color map with modified gradation. Just looking at this color map makes me sick (bitter smile).
Comparing the horizontal direction (year) of the color map shows that it is getting hotter and hotter from the past to the present, and comparing the vertical direction (month) shows that it has been hotter from January to March in recent years. I understand.
This year (2023) July is certainly darker than other years. However, I feel that this year's June is also darker than other years.
With a heat map, it is easy to see the entire data and compare large/small values, The difference in fine color depth can sometimes be visually difficult to discern.
One workaround is to add value labels to the heatmap. "Labels" that can be set on the left side of the graph builder [value label] to display the value in each cell.
It can be clearly seen that this year's June and July were the hottest months of the year. However, from a data visualization perspective, such a chart lacks simplicity and is not recommended as a chart to show to anyone. It's a good idea to remember to display the value labels only for the analyst's confirmation.
Make a line graph
Another countermeasure is to draw a separate line graph. The line graph below is drawn for each year, with the month on the horizontal axis and the temperature difference on the vertical axis. The year has a color gradation, so you can see that the temperature is rising year by year.
Only 2023 is highlighted, but you can clearly see the high temperatures in June and July this year.
Besides that, I am concerned that the red graph is located on the upper side of the vertical axis for almost every month. In recent years, regardless of the month, it is about 1°C hotter than the same month in the past.
Since we are only considering the average temperature of the world here, there should be variations in the heat of each region, The fact that the average temperature of the world as a whole is gradually rising should know.
Hopefully in the next few years it will be time to use the color blue for heatmaps.
by Naohiro Masukawa (JMP Japan)
