---
tags: Master
---

# Data viz bibbia

## Setup R markdown

title: "Italian taste preprocessing"
output:
  html_document:
    toc: yes
    df_print: paged
  html_notebook:
    toc: yes
params:
  echo: yes
  message: no
  warning: no

{r setup, include=FALSE}
knitr::opts_chunk$set(echo = params$echo)
knitr::opts_chunk$set(message = params$message)
knitr::opts_chunk$set(warning = params$warning)
options(reticulate.repl.quiet = TRUE)

## R libraries

library(reticulate) # for accessing Python environment use `py$var_name`
library(tidyverse) # has ggplot + stuff
library(broom) # tidy functions
library(geojsonio) # read geojson
library(umap)     # UMAP
library(smacof)   # MDS
library(Rtsne)    # TSNE
...

## Python libraries

import numpy as np
import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import StandardScaler
from sklearn.manifold import TSNE, MDS
from sklearn.decomposition import PCA
from umap import UMAP

## Fast Excel read

wb = openpyxl.load_workbook("Database generale 2015 2016 2017 TN.xlsx", read_only=True, data_only=True)
sheet = wb["dati"]
rows = sheet.iter_rows(max_row=257, max_col=896, values_only=True)
first_row = next(rows)
data = pd.DataFrame(rows, columns=first_row)

## Read geojson

geo_data <- geojson_read("data/countries.geojson", what = "sp")
geo_data <- tidy(geo_data, region = "ADMIN")

## Maps
https://cengel.github.io/R-spatial/mapping.html#choropleth-mapping-with-ggplot2

## Plot in pandas

data.plot.scatter(x='TSNE1', y='TSNE2')
plt.show()

## Seaborn

import seaborn as sns
...
ax = sns.barplot()
# use matplotlib settings
ax.set(xlabel='common xlabel', ylabel='common ylabel', title='some title')
ax.title("Population Pyramid of the Marketing Funnel", fontsize=22)

## Theme + legend in R
http://www.cookbook-r.com/Graphs/Legends_(ggplot2)
https://ggplot2.tidyverse.org/reference/theme.html
http://zevross.com/blog/2014/08/04/beautiful-plotting-in-r-a-ggplot2-cheatsheet-3/

ggplot(data) +
    # hide aesthetic from legend
    geom_text(data, aes(x, y, label=round(temp), size=4), show_guide=FALSE)
    # modify scales
    scale_y_continuous(breaks = seq(0, 1, by = 0.1), limits = c(0.2, 1)) + # range and tick intervals
    scale_x_continuous(label = function(x){return(paste("My value is", x, "degrees"))}) + # rename ticks labels
    scale_x_date(date_breaks = "3 day", expand = c(0, 0)) + # time ticks and remove empty offset in the axis
    coord_equal() + # if you want same scaling
    # modify theme
    theme_minimal() + # full style (eg.: theme_void(), theme_tufte(),...)
    theme( # overwrite theme settings
        # text
        text = element_text(family = "Verdana"), # all text
        axis.title = element_text(size=20, family="Bauhaus 93", face="bold", margin = margin(10, 0, 10, 0)),
        axis.line.x = element_line(size = 0.3),
        axis.ticks.x = element_line(),
        # background grid
        panel.grid.major.y = element_line(size = 0.3, linetype = "solid", colour = "black"),
        panel.grid.major.x = element_blank(),
        # legend
        legend.position = "top",
        legend.title = element_text(size = 12, colour="chocolate")
    ) +
    # modify text
    labs( 
        title = "",
        subtitle = "",
        caption = "",
        tag = "",
        x = "",
        y = "",
        colour = c("Region" = ""), # rename legend for color
        ... # and more
    )

## Manual shapes/colors

### Manual color

library(tidyverse)
library(ggplot2)
library(Cairo)
library(plotly)

data <- read_csv(file = "economist_data.csv")
cols <- c("Asia Pacific" = "#461313", "Americas" = "#279e27", "East EU Cemt Asia" = "blue", "SSA" = "brown", "EU W. Europe" = "yellow", "MENA" = "grey")
p1 <- ggplot(data, aes(x = CPI, y = HDI, colour = Region)) +
    geom_point(shape = 21, stroke = 1.5, size = 1.5) +
    geom_smooth(
        method = lm,
        aes(
            colour = NULL,
            lty = "diocan"
        ),
        colour = "red"
    ) +
    scale_color_manual(values = cols) +
    scale_fill_discrete(guide = FALSE) +
    theme_classic() +
    theme(
        panel.grid.major.y = element_line(size = 2),
        panel.grid.major = element_line(size = 0.5, color = "black"),
        panel.grid.minor = element_blank()
    ) +
    labs(
        x = "X",
        y = "Y"
    )
p1

### Manual shape

library(ggplot2)
library(tidyverse)
library(ggExtra)

data <- read_csv("iris.csv")

# classic plot :
p1 <- ggplot(data, aes(x = `sepal length`, y = `sepal width`, color = `class`, shape = `class`)) +
    geom_point() +
    scale_shape_manual(values = c("Iris-setosa" = 21, "Iris-versicolor" = 22, "Iris-virginica" = 23)) +
    theme(legend.position = "none")

# marginal density
p2 <- ggMarginal(p1, type = "density", groupColour = TRUE, groupFill = TRUE)

p2

### Manual fill

library(tidyverse)
library(ggplot2)
library(Cairo)
library(plotly)
library(gridExtra)

data <- read_tsv(file = "barplot_data.csv")
data$SEQC_SIGN <- as.character(sign(data$FoldChange_SEQC))
data$NANO_SIGN <- as.character(sign(data$FoldChange_Nanostring))

p1 <- ggplot(data, aes(x = abs(FoldChange_SEQC), y = Gene)) +
    geom_bar(stat = "identity", aes(fill = SEQC_SIGN), show.legend = FALSE) +
    geom_text(aes(label = strrep("*", floor(-log(Pvalue_SEQC) / 8)))) +
    scale_fill_manual(values = c("-1" = "red", "1" = "green")) +
    theme_minimal() +
    theme(axis.text.y = element_text(hjust = 0.5))

p2 <- ggplot(data, aes(x = -abs(FoldChange_Nanostring), y = Gene)) +
    geom_bar(stat = "identity", aes(fill = NANO_SIGN), show.legend = FALSE) +
    geom_text(aes(label = strrep("*", floor(-log(Pvalue_SEQC) / 8)))) +
    scale_fill_manual(values = c("-1" = "red", "1" = "green")) +
    scale_x_continuous(
        labels = function(x) signif(abs(x), 3),
        breaks = c(0, -2, -4, -6, -8),
        limits = c(-9, 0)
    ) +
    theme_minimal()

grid.arrange(p2, p1, ncol = 2)

## Data transform in r

# set new column as "-1" or "1" (need categorical, not numeric)
data$SEQC_SIGN <- as.character(sign(data$FoldChange_SEQC))
data$NANO_SIGN <- as.character(sign(data$FoldChange_Nanostring))

# unique
countries <- unique(data$Country)

## Faceting in R

p1 <- ggplot(data, aes(x = abs(FoldChange_SEQC), y = Gene)) +
    geom_bar(stat = "identity", aes(fill = SEQC_SIGN), show.legend = FALSE) +
    geom_text(aes(label = strrep("*", floor(-log(Pvalue_SEQC) / 8)))) +
    scale_fill_manual(values = c("-1" = "red", "1" = "green")) +
    theme_minimal() +
    theme(axis.text.y = element_text(hjust = 0.5))

p2 <- ggplot(data, aes(x = -abs(FoldChange_Nanostring), y = Gene)) +
    geom_bar(stat = "identity", aes(fill = NANO_SIGN), show.legend = FALSE) +
    geom_text(aes(label = strrep("*", floor(-log(Pvalue_SEQC) / 8)))) +
    scale_fill_manual(values = c("-1" = "red", "1" = "green")) +
    scale_x_continuous(
        labels = function(x) signif(abs(x), 3),
        breaks = c(0, -2, -4, -6, -8),
        limits = c(-9, 0)
    ) +
    theme_minimal()

grid.arrange(p2, p1, ncol = 2)


europe <- mutate(europe, gud = (HDI > 0.4))