Jose V. Die

Unix

get the radio show length (in seconds)

mdls *.mp3
mdls *.mp3 | grep kMDItemDurationSeconds
mdls *.mp3 | grep kMDItemDurationSeconds | cut -d '=' -f 2
mdls *.mp3 | grep kMDItemDurationSeconds | cut -d '=' -f 2 | cut -c 2-
mdls *.mp3 | grep kMDItemDurationSeconds | cut -d '=' -f 2 | cut -c 2- > duracion.csv

get the name of each episode

mdls *.mp3 | grep kMDItemDisplayName
mdls *.mp3 | grep kMDItemDisplayName | cut -d '=' -f 2
mdls *.mp3 | grep kMDItemDisplayName | cut -d '=' -f 2 | cut -c 2-
mdls *.mp3 | grep kMDItemDisplayName | cut -d '=' -f 2 | cut -c 2- > episodios.csv

R

Dependencies

library(stringr) 
library(lubridate)
library(ggplot2)
library(dplyr)

Read the file with the radio show length (in seconds)

segundos <- readLines("duracion.csv")

Read the file with the radio show date

fechas <- readLines("episodios.csv")

Check both files have the same elements

length(segundos) == length(fecha)

Split each element based on a pattern

tmp <- str_split(fecha, "-cvradio")

tmp <- unlist(tmp)

Three episodes require manually edition.
The date for each episode can be found in the A Ciencia Cierta podcast website .

tmp[1] = "\"acienciacierta01122016"
tmp[3] = "\"acienciacierta22122016"
tmp[9] = "\"acienciacierta19012017"

Now, we select only those elements containing the date based on a pattern

str_subset(tmp, "acienciacierta")

The date is contained in the last eight elements

str_sub(str_subset(tmp, "acienciacierta"), -8, -1)

We can use the parser dmy from the lubridate package

fechas <- dmy(str_sub(str_subset(tmp, "acienciacierta"), -8, -1))

Simple plot on weekdays when the radio show is on

qplot(weekdays(fechas))

Finally, we make a new dataset, so we can use it in ggplot2

dat <- tibble(fecha = fechas, 
       minutos = as.numeric(segundos / 60) 
       
dat %>% ggplot(aes(x = fecha, y = minutos)) +
  geom_point() + 
  geom_smooth(method = "lm", col = "steelblue") + 
  ggtitle("A ciencia cierta", "Antonio Rivera")

A regression linear model :

my_model = lm(data = dat, minutos~fecha)
summary(my_model)
# coefficients(my_model)