Scatterplot with scaled markers and color values (not finished)

This article is not finished yet.

Before starting this you should be familiar with the 2 axis scatterplot. This article will add two more variables in the mix so we will have X and Y axis, and the markers scale and color will represent two more variables.

The data we will use in this article is population data for norwegian counties. Every county has the following variables. English meaning of variable names behind the // comments

       "01 Østfold";"2019K2";"population";297898
        "01 Østfold";"2019K2";"born";691
        "01 Østfold";"2019K2";"dead";628
        "01 Østfold";"2019K2";"Fødselsoverskot";63 //Birth surplus
        "01 Østfold";"2019K2";"Innvandring";467  //Immigration
        "01 Østfold";"2019K2";"Utvandring";218   //Emmigration
        "01 Østfold";"2019K2";"Innflytting, innalandsk";2203  //Moving to county from within norway
        "01 Østfold";"2019K2";"Utflytting, innalandsk";1370   //Moving from county to within norway
        "01 Østfold";"2019K2";"Nettoinnflytting, inkl. inn- og utvandring";1082  //Net moving in
        "01 Østfold";"2019K2";"Folkevekst";1145   //population growth
        "01 Østfold";"2019K2";"Befolkning ved utgangen av kvartalet";299043  //population by end if quarter

Of these variables I have chosen births, deaths, population and a calculated value of birth surplus per 100000 people. Birth surplus means the amount of births minus the amount of deaths. Here is the scatterplot:

X axis: Deaths. Y axis: Births. Scale: Population. Color: Birth surplus per 100000 people

You can download the files here

Compared to the other scatterplot this one has simpler data binding as the csv file is smaller. On the other hand it goes a bit further in encoding the values to various visual elements. It uses two axis (X & Y both linear), it uses colour from a colour scale and it encodes population to the area of the circles. There are many ways to visualize a data set like this and you might want to choose other variables. For the scaled markers I reccommend using population in this case but other variables can work just fine depending on the project. You can resize the browser to see how the chart adapts to the various widths while maintaining the height. This is a choice I have made based on experience. The most common way to make a chart responsive is probably to scale it on both axis. I find this often results in the chart being less readable since fonts and axis are normally also scaled down to a very small size. The horizontal only scaling maintains the height and font sizes. Every bubble in the scatterplot is moved individually but the scale of the bubbles are maintained.