data mountains¶

turn your data into mountains!¶

census 2021 data for england and wales is out now¶

we want to show the data on a map¶

...why? most council data tends to be geospatial as the work always happens somewhere (also 60% of all data is geospatial, allegedly, so maps are important for datavis)¶

i am interested in the part of england that has the the fastest growing population and the highest population density:¶

The London Borough of Tower Hamlets¶

where is tower hamlets?¶

image.png

we use choropleth maps a lot¶

but this one looks washed out...¶

image.png

...oh yes, the data is not normally distributed¶

image.png

so we log transform the data¶

...but we lose the values in the legend¶

image.png

let's try a density equalising cartogram¶

shout to the r cartogram package¶

nice, but harder to overlay on other data¶

posit

image.png

...remember this?¶

image.png

...ah yes, that is a very small census area, so the density is high relative to the space¶

image.png

here it is, a statistical geography that is just one building¶

... it happens - this is why the density is so high, that pointy datavis was a good way to see this (instead of choropleth or cartogram)¶

image.png

playing with datawrapper was fun¶

the pointy map symbols reminded me of middle earth¶

image.png

i tried to emulate hand drawn maps in python¶

the essence of data mountains is a simple function turns a point into a mountain, sized by a data attribute of that point¶

function - project to spatial

population density¶

pop-density-2021-lb-tower-hamlets.png

population density is a bit odd at this spatial resolution¶

pop-density-2021-lb-tower-hamlets.png

let's try another dataset¶

one bedroom homes¶

1 bed homes

two bedroom homes¶

2 bed homes

three bedroom homes¶

3 bed homes

four or more bedroom homes¶

4+ bed homes

data mountains in a more central area¶

westminster population density¶

westminster population density

data mountains in an inner-outer city area¶

westminster population density

made with #nbdev!¶

using nbdev for this project made it extra fun, it helps you show good coder practice¶

  1. it cleans cell metadata during version control, avoiding merge conflicts
  2. it enables you to easily share:
    1. code
    2. docs
    3. tests
    4. modules ... all at once!

...finally... shout to jarek (sojourner.rocks), ed (wikidata osm link), mark, lee, and all the council coders

council coders