gibsramen

import matplotlib as mpl
from matplotlib import pyplot as plt
from matplotlib.patches import Patch
import numpy as np
import pandas as pd
import seaborn as sns


# Paper 1
paper_1_start = "02/05/2018"

gibsramen

from typing import Tuple
import warnings

import biom
import click
from gemelli.rpca import rpca
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
import seaborn as sns

gibsramen

library(biomformat)
library(ANCOMBC)
library(phyloseq)

tbl_file <- "data/shi_age_prediction/processed/processed_skin_tbl.biom"
tbl <- biomformat::read_biom(tbl_file)
tbl <- as.data.frame(as.matrix(biomformat::biom_data(tbl)))

md_file <- "data/shi_age_prediction/processed/processed_skin_md.tsv"
md <- read.table(md_file, sep="\t", row.names=1, header=T)

gibsramen

get_ord_dataframe <- function(lines, linestart){
  header_line <- strsplit(lines[linestart], split="\t")[[1]]
  num_rows = as.numeric(header_line[2])
  num_cols = as.numeric(header_line[3])
  if (num_rows == 0){return(data.frame())}
  
  data <- strsplit(lines[(linestart+1):(linestart+num_rows)], split="\t")
  names <- unlist(lapply(data, function(x) x[1]))
  coords <- lapply(data, function(x) strsplit(x[2:(2+num_cols-1)], split="\t"))
  coords <- data.frame(matrix(unlist(coords), nrow=length(coords), byrow=T), stringsAsFactors=FALSE)

gibsramen

def newton_method(f, df, x0, err, W):
    """
    Estimate p using Newton-Raphson method.
    
    Output: root approximation given error criterion (err)
    """
    delta = err+1
    x = x0
    while delta > err:
        x1 = x - f(x, W) / df(x)

gibsramen

x <- 1:10
y <- 1 + 2.5 * x
N <- length(x)
x_mean <- mean(x)
y_mean <- mean(y)
m <- sum((x - x_mean) * (y - y_mean)) / sum((x - x_mean)^2)
b <- y_mean - m * x_mean

# m = 2.5
# b = 1

gibsramen

> summarize(by_sex, median(LIMIT_BAL))
# A tibble: 2 × 2
    SEX `median(LIMIT_BAL)`
  <chr>               <dbl>
1     1              130000
2     2              150000
> summarize(by_sex, length(dflt[dflt=='1'])/length(SEX))
# A tibble: 2 × 2
    SEX `length(dflt[dflt == "1"])/length(SEX)`
  <chr>                                   <dbl>

gibsramen

library(dplyr)

d <- read.csv('default of credit card clients.csv', header=F, stringsAsFactors=F)
colnames(d) <- unlist(d[2,])
d <- d[-c(1,2),]
colnames(d)[25] <- 'dflt'
d$LIMIT_BAL <- as.numeric(d$LIMIT_BAL)

# MALE VS. FEMALE
# ----------------

gibsramen

library(ggplot2)

votes <- read.csv('csv-ballot-measures.csv', header=TRUE, stringsAsFactors=FALSE)
votes <- votes[,-1:-2]
d.votes <- votes[which(votes$BALLOT_MEASURE_ID==62),]

drops <- c('COUNTY_ID', 'BALLOT_MEASURE_ID', 'BALLOT_MEASURE_NAME', 'BALLOT_MEASURE_TITLE')
d.votes <- d.votes[, !names(votes) %in% drops]

d.votes$YES_COUNT <- as.numeric(gsub(',', '', d.votes$YES_COUNT))