Graph of temperature changes for the year in the burg

[Kerravitarr]

I'm bad at writing in English. Please excuse me for this.
It seems illogical to me the parameter of the average temperature for the year in the burg.
Therefore, I found a database of meteorological measurements at https://ru.climate-data.org/
And I analyzed it. With the help of a neural network. I approximately established the dependence of Dy[X] and Dd[X], Md[X] = f(t,p,l).
Dy[X] - how much the temperature in winter differs from the temperature in summer
Md[X] - how much the temperature during the day differs from the temperature at night, on average per year
Dd[X] - how much the temperature difference between day and night is not constant in winter and summer
t - average temperature for the year
p - average rainfall for the year
l - latitude
Now, for each city, you can plot the temperature change. Based on t,p and l.

ATTENTION!!!
The characteristics of the planet and its orbit coincide with those of the earth!

Further, I will use the Russian language, if you want to know my reasoning, please use the Google Translate capabilities.

Я изменил свою первоначальную идею и доработал её. Что касается качества и используемых функций - жду критики. JavaScript - не знакомый мне язык и я не уверен, что выполнил всё верно.
Что касается самой идеи - её реализация пока далека от совершенства. Я понял, что если установить в настройках мира температуру на экваторе ниже, чем на Земле, то графики получаются слишком далекими от реальности. Эта ошибка связана с тем, что Dy[X] зависит от широты, а не среднегодовой температуры. К сожалению добиться хорошей сходимости удалось только расширением нейросети - ещё больше магических чисел.
Теперь, на основе прошлой выборки из 899 городов, попущенных через нейросеть миллион раз среднеквадратичная ошибка стала 3%.
Входными данными для сети являются широта и влажность, а средняя температура иcпользуется только как ориентир при построении графиков.

[Azgaar]

The idea is definitely great and benefit the project.

I would like to request to add a legend as the color-coding is not 100% clear. Other changes (clean up, language and js fixes) I can do on my own after the merge.

[Kerravitarr]

I added a legend.
I hope to bring more to the project in the future.

[Kerravitarr]

I decided to test my theory on my map and noticed a big discrepancy.
Latitude greater than 0 and less than 0 does not equally affect Dy[X] and Dd[X], Md[X]. It seems to me that this is a mistake. It could appear due to lack of data.
Perhaps in the future, it will be possible to expand the number of input parameters of the neural network. Add the ratio between the temperature in the burg and the average temperature of the planet.

[Kerravitarr]

I have updated the data again. Now the latitude sign does not affect the result.

There is still a problem that worries me. In my opinion, the graph should depend on the height too. But so far I don't have any data.

[Azgaar]

Are you going to continue improving it?

[Kerravitarr]

Maybe in the future.
I get too carried away sometimes, but now I'm happy with the result.

[Azgaar]

Ok, I will merge when I have time

[Azgaar]

Merging to a separate branch. I will do required fixes and then merge to master.

[Azgaar]

Merged to master as a part of #734. I have refactored the code and UI a bit, but the idea stood intact.

Thanks @Kerravitarr for creating!

[Azgaar]

Deployed, see also https://www.patreon.com/posts/62171949

[Kerravitarr]

Thanks!
I found another small improvement. Line 51 in modules/ui/temperature-graph.js
Instead of
const yearDelTmpSig = (lstOut[1]*13.541688670361175+0.1414213562373084) > yearSig ? yearSig : (lstOut[1]*13.541688670361175+0.1414213562373084);
, you can write
const yearDelTmpSig = Math.min(lstOut[1]*13.541688670361175+0.1414213562373084, yearSig);

I don't know much about the climate. Therefore, I don’t know what rains in the city, clouds, etc. depend on. But if it becomes important for my story, maybe I'll figure it out.

[Azgaar]

Hello. You can create a new PR. As for the weather conditions, we have only one value - precipitation. That can be used the same way as you have utilized the temperature.