nacnudus

# Inspired by http://www.brodrigues.co/blog/2018-09-11-human_to_machine/
# https://twitter.com/brodriguesco/status/1039604517287931904
# "You can find the data I will use here. Click on the “Time use” folder and you can download the workbook."
# http://statistiques.public.lu/stat/ReportFolders/ReportFolder.aspx?IF_Language=eng&MainTheme=3&FldrName=1&RFPath=14306

library(tidyverse)
library(tidyxl)
library(unpivotr)
library(lubridate)

nacnudus

# Inspired by http://www.brodrigues.co/blog/2018-09-11-human_to_machine/
# https://twitter.com/brodriguesco/status/1039604517287931904
# "You can find the data I will use here. Click on the “Time use” folder and you can download the workbook."
# http://statistiques.public.lu/stat/ReportFolders/ReportFolder.aspx?IF_Language=eng&MainTheme=3&FldrName=1&RFPath=14306

# This time using experimental unpivotr code to allow custom filtering of header cells, rather than having to reposition them.
# https://github.com/nacnudus/unpivotr/commit/0961ec3c3e17b34755f0fce94db7f5bf380d43ce

library(tidyverse)
library(tidyxl)

nacnudus

library(tidyverse)
library(ompr)
library(ompr.roi)
library(ROI.plugin.glpk)

M <- 3 # Volunteers (rows)
N <- 4 # Jobs (combination of role at given time and location) (columnss)

# Jobs are:
# 1. Greet (8am-9am)

nacnudus

# ---------------------------------------
# untidy data 
# ---------------------------------------

# this dataset is a sample of the kind of data that might appear in the wild, 
# particularly when dealing with government data, 
# particularly when trying to convert an output table or individual report 
# back into a raw data format that can be analyzed. 

# in these test datasets, discrete observations are spread out across multiple rows. 

nacnudus

# Compare performance of ls() with as.list.environment() for counting objects
# length(as.list.environment()) is fastest

env <- environment(plot) # A handy, large environment

bench::mark(length(ls(env)), length(as.list(env)))
# # A tibble: 2 x 13
#   expression                min   median `itr/sec` mem_alloc `gc/sec` n_itr  n_gc total_time result    memory time     gc
#   <bch:expr>           <bch:tm> <bch:tm>     <dbl> <bch:byt>    <dbl> <int> <dbl>   <bch:tm> <list>    <list> <list>   <list>
# 1 length(ls(env))        4.69ms   4.96ms      198.        NA        0    99     0      501ms <int [1]> <NULL> <bch:tm> <tibble [99 × 3]>

nacnudus

library(tidyverse)
library(tidyxl)
library(unpivotr)

#' Convert relevant character values to dates
#'
#' @param cells Data frame derived from `tidyxl::xlsx_cells()` or
#' `readr::melt_csv()` or similar.
#' @param condition An expression that returns a logical value,
#' is defined in terms of the columns in `cells`.  Similar to `dplyr::filter()`.

nacnudus

# Script to analyse Gmail message sizes by from/to/subject

# 1. Download Gmail from Google Takeout, specifically the folders "Inbox",
# "Archived", "Sent", and "Bin".
# 2. Extract it into the working directory.  It should create the folders
# Takeout/Mail
# 3. Run this script in the working directory

library(tidyverse)
library(tm.plugin.mail)

nacnudus

# The fastest implementation available is in mlpack (C++)
using Pkg
Pkg.add("mlpack")
using mlpack
x = rand(5,2)
x_emst = mlpack.emst(x)

# A far slower implementation is in EMST.jl.  As of 2022-03-14 my fork has been updated to run with Julia v0.7+using Pkg
using Pkg
Pkg.add(url="https://github.com/nacnudus/EMST.jl", rev="julia-v0.7")

nacnudus

#!/bin/bash

ENTRY=$(/usr/bin/yad --title "Activity log" --text="What are you doing?" --entry --auto-kill)
DATETIME=$(date --iso-8601=seconds)

[[ ! -z "$ENTRY" ]] && echo -e "${DATETIME}\t${ENTRY}">> /home/nacnudus/gds/log.log

# Schedule in crontab Every six minutes between 9am and 6pm on weekdays
# */6 9-18 * * 1-5 env DISPLAY=:0 && /usr/bin/bash /path/to/this_file.sh

nacnudus

library(igraph)
library(tidygraph)

# Create an igraph
# c - d
#   - e
#   - f - b
ig <- graph_from_literal(f-+b, c-+e:f, c-+d)
ig