christophergandrud

Bash script to create a skeleton Streamlit app

To run the script from the command line use:

curl -s https://gist.githubusercontent.com/christophergandrud/8f467beaf72a0ae1f417c3e8364c0039/raw/774b35818c72575e42aa703f472ee43cbc9d4c79/streamlit-build.sh -o script.sh && chmod +x script.sh && ./script.sh && rm script.sh

christophergandrud

"""
Create the card game "War" 
where the player with the most cards at the end of 1 round wins
(written using GitHub Copilot)
"""

# Create a range with max from len(x)
def range_len(x) :
    """
    Return the range of the length of x

christophergandrud

#' Multi-core replicate. From the rethinking package:
#' <https://github.com/rmcelreath/rethinking/blob/3b48ec8dfda4840b9dce096d0cb9406589ef7923/R/utilities.r#L206>
#' 
#' @param n integer: the number of replications.
#' @param expr the expression (a language object, usually a call) to evaluate repeatedly.
#' @param refresh status update refresh interval
#' @param mc.cores number of cores to use
#' 
#' @importFrom parallel mclapply
#' @export 

christophergandrud

"""
Functions for recursively calculating mean and variance and comparing
    comparing to full sample calculated mean and variance.

Functions based on:

- Chou (2021, 5) <https://arxiv.org/pdf/2102.03316.pdf>

- Chan et al (1983) <http://www.cs.yale.edu/publications/techreports/tr222.pdf>
"""

christophergandrud

# ---
# title: "Covid-19 Vaccination Program Growth Rates"
# author: "Christopher Gandrud"
# date: "2021-02-21"
# ---

# For hrbragg installation instructions see:
# https://git.rud.is/hrbrmstr/hrbragg#installation
xfun::pkg_attach2("hrbragg", "tsibble", "tidyverse")

christophergandrud

"""
	robustness_value(;fit::StatsModels.TableRegressionModel, treatment::String = "D", q::Int64 = 1)

Find the Cinelli and Hazlett (2020) robustness value for a generalised linear model. 
The response will be a in the range 0 and 1. Values closer to 0 indicate that 
the conclusions from the original model are highly subject to omitted variable 
bias. Values closer to 1 indicate that the original model is less likely to be 
caused by omitted variable bias.
"""
function robustness_value(;fit::StatsModels.TableRegressionModel,

christophergandrud

"""
    ebt_sampler(total_questions::Int = 310)
Create question set to practice for the Deutscher Einbürgerungstest. 
The official question set can be found at: <http://oet.bamf.de/pls/oetut/f?p=514:1:329473569276328:::::>.
The question set has 310 questions, but who knows, maybe this could change. 
Adjust the total question set with `total_questions::Int`.
"""
function ebt_sampler(total_questions::Int = 310) 
    i::Int = -9; u::Int = 0
    out = zeros(Int, 0)

christophergandrud

# Create lag and difference variables for a grouped time series (grouped by a variable called "country")
#
# @param df a data frame containing the variable to lag and lead
# @param x a column name in `df` to lag and lead
# @param lag integer period to lag and difference
#
# @importFrom tsibble difference
# @importFrom dplyr %>% lag

create_lags_diffs <- function(df, x, lag = 1) {

christophergandrud

library(xfun)
pkg_attach2("tidyverse", "poweRlaw", "furrr")
plan(multiprocess)


# Simulate from a discrete power law
# @param counter numeric, to enable mapping
# @param n integer, number of draws
# @param xmin numeric, minimum value subject to power law
# @param alpha, scaling parameter

christophergandrud