derekpowell

make_explainer_obj <- function(fitted_workflow){
  fitted_model <-
    fitted_workflow %>% 
    extract_fit_parsnip() # <- parsnip model_fit object

  feature_data <- 
    fitted_workflow %>% 
    extract_mold() %>% 
    pluck("predictors") 
  

derekpowell

# script to create sqlite database of daily GHCN weather station data
# adding a table of US zipcodes for my own purposes (easily commented out)
# cribs heavily from https://github.com/dylburger/noaa-ghcn-weather-data

# Settings

import sqlalchemy
import os
import urllib.request
import pandas as pd

derekpowell

cumulative_intercept_prior <- function(k, sd = 2, alpha = 1, beta = 1,
                                      shape = c("flat", "middle", "rightskewed", "leftskewed")) {
  ## Creates priors on intercepts for cumulative() family regression.
  ## Assumes that probability of response options follow cumulative beta 
  ## distribution specified by a and b or by "shape" argument.
  ##
  ## k = number of categories
  ## sd = std dev of normal over intercept
  ## a, b = alpha and beta specifying shape of distribution (defaults to uniform)
  ## shape = string specifying pre-defined distribution shape 

derekpowell

def horseshoe_prior(name, X, y, m, v, s):
    '''
    Regularizing horseshoe prior as introduced by Piironen & Vehtari
    https://arxiv.org/pdf/1707.01694.pdf
    
    name: variable name
    X: X (2-d array)
    y: y (for setting pseudo-variance)
    m: expected number of relevant features (must be < total N)
    v: regularizing student-t df

derekpowell

def gather( df, key, value, cols ):
    id_vars = [ col for col in df.columns if col not in cols ]
    id_values = cols
    var_name = key
    value_name = value
    return pd.melt( df, id_vars, id_values, var_name, value_name )

def spread( df, index, columns, values ):
    return df.pivot(index, columns, values).reset_index(level=index).rename_axis(None,axis=1)

derekpowell

cbrm <- function(formula,
                 data,
                 family = gaussian(),
                 prior = NULL,
                 autocor = NULL,
                 cov_ranef = NULL,
                 sample_prior = c("no", "yes", "only"),
                 sparse = FALSE,
                 knots = NULL,
                 stan_funs = NULL,

derekpowell

rescale_beta <- function(x, lower, upper) {
  # rescales onto the open interval (0,1)
  # rescales over theoretical bounds of measurement, specified by "upper" and "lower"

  N <- length(x)
  res <- (x - lower) / (upper - lower)
  res <- (res * (N - 1) + .5) / N

  return(as.vector(res))
}

derekpowell

read_qualtrics_csv <- function(fname) {
  headers <- as.matrix(read.csv(fname, skip = 0, header = F, nrows = 1, as.is = T))
  df <- read_csv(fname, skip = 3, col_names = headers)
  
  df <- df %>% 
    filter(DistributionChannel=="anonymous")
  
  return(df)
}

derekpowell

model_name <- brm(
  DV ~ formula,
  data = d,
  family = normal(), # student(), #cumulative(), #bernoulli(), etc
  control = list(adapt_delta = .80),
  cores = parallel::detectCores(),
  iter = 2000)

derekpowell

# Author: Derek Powell
# Date: 8/30/18, 4:49 PM
# ---
# Script to redact workerIds and ip addresses from qualtrics files.
# Script looks for "date_private/" directory, saves resulting data in "data" directory.
# Personal info is replaced with a "hash" using xxhash64,
# a super fast hash algo w/ short resulting hashes (confirmed appropriate for this use)

suppressMessages(library(tidyverse))
suppressMessages(library(digest))