library(tidyverse)
library(mosaic) # Our all-in-one package
library(skimr) # Looking at data
library(ggformula) # Our plotting package
library(visdat) # Mapping missing data
library(naniar) # Missing data visualization and munging
library(janitor) # Clean the data
library(tinytable) # Printing Tables for our data
library(DT) # Interactive Tables for our data
#
# To install packages fro github when not available on CRAN
# library(pak)
# pak::pak("rpruim/Lock5withR")
library(Lock5withR)
library(Lock5Data) # Some neat little datasets from a lovely textbook
Charts and How they are generated from Data
2021-11-01
““He is one of those who don’t want millions, but an answer to their questions.”
— Alyosha, in The Brothers Karamazov
Some Pictures
- It helps sift facts from mere statements: for example:
- Visuals are a good starting point to make hypotheses of what may be happening in the situation represented by the data
Recap: What is Tidy Data?
Let us recall first what we meant by tidy data:
How would an Art + Design student look at a Data Visualization? Like a painting, or an art work!!
Look at these paintings from Kandinsky. What’s the first thing you would notice? From a cultural standpoint: the shapes, colours, textures…all things geometrical are the first thing you would notice.
Kandinsky used geometric shapes to convey ideas like emotions, relationships, movement, tension, harmony, chaos, balance, contrast and so on. These ideas are what we find encoded in data, in Qual and Quant variables!!
So how do we become Kandinsky-with-a-laptop?
So first:
- How many geometric things do we know?
- Shapes? Lines? Axes? Curves? Angles? Patterns? Textures? Colours? Sizes? Positions? Lengths? Heights? Breadths? Radii? Textures?
- All these are geometric aspects or aesthetics, each with a unique property.
- Some “geometric things” which we might consider are shown in the figure below.
A Data Visualization Example
Data Viz = Data + Geometry
Facets
- Finally, if the graph is too busy, with lots of colours and shapes, then we can split the graph into many small multiples or facets, each showing a subset of the data.
- This is called faceting and is a powerful way to reduce cognitive load on the viewer.
Code
set.seed(1947)
diamonds %>%
slice_sample(n = 150, weight_by = cut) %>%
gf_point(price ~ carat | clarity,
colour = ~cut,
shape = ~cut,
size = 2, data = .
) %>%
gf_labs(
title = "Plot Title = DIAMONDS ARE FOREVER",
subtitle = "Plot Subtitle = AND A GIRL'S BEST FRIEND",
caption = "Plot Caption = From the diamonds dataset",
x = "x-Axis Title = CARAT",
y = "y-Axis Title = PRICE"
) %>%
# Use same name for scales to merge legends
gf_refine(
scale_color_brewer(
name = "Legend = DIAMOND QUALITY",
palette = "Set1"
),
scale_shape_manual(
name = "Legend = DIAMOND QUALITY",
values = c(15:21)
)
)
Data Science Workflow
