ledell

# How to get k-fold metrics for all the H2O AutoML models in R
# Adapted from: http://docs.h2o.ai/h2o/latest-stable/h2o-docs/automl.html

library(h2o)

h2o.init()

# Import a sample binary outcome train/test set into H2O
train <- h2o.importFile("https://s3.amazonaws.com/erin-data/higgs/higgs_train_10k.csv")
test <- h2o.importFile("https://s3.amazonaws.com/erin-data/higgs/higgs_test_5k.csv")

jnolis

#     _____       __                          __
#    |__  /      / /__  ____ _____ ____  ____/ /
#     /_ <______/ / _ \/ __ `/ __ `/ _ \/ __  / 
#   ___/ /_____/ /  __/ /_/ / /_/ /  __/ /_/ /  
#  /____/     /_/\___/\__, /\__, /\___/\__,_/   
#                    /____//____/               
#         RTWEET + 3-LEGGED-AUTH DEMO (LITE)

# This code demonstrates how to do 3-legged authentication for Twitter
# using the {rtweet} package. Based heavily on code from Michael Kearney.

eoppe1022

library(tidyverse)
library(rtweet)
library(markovifyR)

## Get tweets with "rtweet"
## FYI you need to set up "rtweet" beforehand
yost_data <- get_timeline("@travisyost", n = 10000)

## Take out RTs and create Markov model
markov_model <- yost_data %>%

Dpananos

library(tidyverse)

n.samples = 25

set.seed(5)
rerun(20,rnorm(n.samples)) %>% 
  map_dfr(~data_frame(data = list(.x)), .id = 'samples') %>% 
  mutate(mu = map_dbl(data,mean), 
         se = map_dbl(data,~sd(.x)/sqrt(length(.x))), 
         top = mu +1.96*se,

DavisVaughan

library(furrr)

# localhost -> AWS EC2 linux -> Docker running on that -> R

# dm_create() and dm_ip() are from an unreleased R pkg I whipped up, dockermachinery
# https://github.com/DavisVaughan/dockermachinery

# Creates 1 t2.micro EC2 instance
dm_create("amazonec2", "dockertest")

khakieconomics

library(rstanarm); library(tidyverse)
options(mc.cores = parallel::detectCores())
set.seed(42)
data("radon")
head(treatment_sample)

# Some levels have no variance in the outcomes, making likelihood estimates impossible
# Adding a tiny bit of noise fixes the problem
radon$log_uranium <- rnorm(nrow(radon), radon$log_uranium, 0.05)

thomasp85

library(future)
trim_model <- function(model, predictor = predict, ..., ignore_warnings = TRUE) {
  # Cache the correct output
  true_pred <- predictor(model, ...)
  
  # Treat prediction warnings as errors?
  if (!ignore_warnings) {
    old_ops <- options(warn = 2)
    on.exit(options(old_ops))
  }

DavisVaughan

// [[Rcpp::depends(xtensor)]]

#include <numeric>
#include "xtensor/xmath.hpp"
#include "xtensor-r/rarray.hpp"
#include <Rcpp.h>

using namespace Rcpp;

// [[Rcpp::plugins(cpp14)]]

noamross

library(rtweet) #rtweet API creds should already be set up
library(stringi)
library(dplyr)

friends = get_friends(user="noamross")
followers = get_followers("noamross")
tweeps_id = distinct(bind_rows(friends, followers))
tweeps_info = lookup_users(tweeps_id$user_id)

# A regex for a visit to Durham

mike-lawrence

functions{
	// GP: computes noiseless Gaussian Process
	vector GP(real volatility, real amplitude, vector normal01, int n_x, real[] x ) {
		matrix[n_x,n_x] cov_mat ;
		real amplitude_sq_plus_jitter ;
		amplitude_sq_plus_jitter = amplitude^2 + 1e-6 ;
		cov_mat = cov_exp_quad(x, amplitude, 1/volatility) ;
		for(i in 1:n_x){
			cov_mat[i,i] = amplitude_sq_plus_jitter ;
		}