thoolihan

from datetime import datetime

def get_curr_time():
    return datetime.now().strftime("%Y.%m.%d.%H.%M.%S")

def get_start_time():
    return _start_time if _start_time else get_curr_time()

_start_time = get_start_time()

thoolihan

library(ggplot2)
library(dplyr)
library(R.utils)

teams <- data.frame(team = c('warriors', 'rockets',
                              'cavaliers', 'celtics'),
                    odds_nw = c(5,9,8,20),
                    odds_w = c(9,4,1,1))

# raw probabilities sum to more than 1 because of house take

thoolihan

import numpy as np
import pandas as pd
from sklearn.base import BaseEstimator, TransformerMixin

# Adapted from https://www.kaggle.com/ogrellier/python-target-encoding-for-categorical-features

class TargetEncoder(BaseEstimator, TransformerMixin):
    def __init__(self, columns, noise_level = 0):
        self.columns = columns
        self.maps = {}

thoolihan

> df <- data.frame(date = c('2017-10-01', '2017-10-11'))
> df$date <- as.Date(df$date)
> df
        date
1 2017-10-01
2 2017-10-11
> sapply(df$date, function(d) {if (d < as.Date('2017-10-07')) 1 else 0})
[1] 1 0
> df$week1 <- sapply(df$date, function(d) {if (d < as.Date('2017-10-07')) 1 else 0})
> df

thoolihan

library(tidyverse)

mtcars %>%
  mutate(car = rownames(.)) %>%
  arrange(hp) %>%
  ggplot(aes(x = car, y = hp)) +
  geom_point() + 
  theme(axis.text.x = element_text(angle = 45, hjust = 1))

thoolihan

# simulating https://en.wikipedia.org/wiki/Markov_chain#/media/File:Markovkate_01.svg
library(ggplot2)

means = c()

ntimes <- 1000

for (t in 1:ntimes) {
  n <- 1000
  state <- c(1)

thoolihan

library(dplyr)

n <- 1000000
data <- data.frame(id = 1:n,
                   red = sample(0:255, size = n, replace = TRUE),
                   green = sample(0:255, size = n, replace = TRUE),
                   blue = sample(255, size = n, replace = TRUE))

query <- list(red = 80, green = 90, blue = 255)
                             

thoolihan

library(dplyr)
library(ggplot2)

doors <- 1:3
sample_doors <- function() { return(sample(doors, size = 1000, replace = TRUE))}
games <- data.frame(prize = sample_doors(), pick = sample_doors())

games$strategy <- factor(ifelse(games$prize == games$pick, 'stay', 'switch'))

monte_show <- function(prize, pick) {

thoolihan

library(ggplot2)

cars <- mtcars
cars$cyl <- factor(cars$cyl, labels = 
                     c('Four cylinder', 'Six cylinder', 'Eight cylinder'))

features <- c('wt', 'qsec')
n_clusters <- 3
car_clusters <- kmeans(cars[, features], n_clusters, nstart = 30)

thoolihan

import numpy as np

a = np.arange(1000)
a = a.reshape(2, 500)
a = a.resize((2,600), refcheck = False)