karma.Rmd

---
title: 'Overview of todo performance alla Todoist Karma'
output:
  rmdformats::html_clean:
    number_sections: false
    theme: paper
    highlight: default
    dev: svg
    df_print: kable
    code_folding: show
    fig_width: 16
    self_contained: true
    gallery: true
---

```{r set-options-read-chunks, echo=FALSE, message=FALSE, warning=FALSE, bootstrap.show.code = FALSE}
knitr::opts_chunk$set(results = 'markup', echo=TRUE, message=FALSE, warning=FALSE, dpi = 600,
                      fig.asp = 0.618)
options(width=120)
library(tidyverse)
library(ggplot2)
library(patchwork)
library(lemmens)
devtools::load_all()
palette <- 'Dark2'
```

```{r load-data}
loc <- '~/nextcloud/todo/'
tasks <- load_tasks(file.path(loc, 'todo.txt'))
done <- load_tasks(file.path(loc, 'done.txt'))
done_old <- load_tasks(file.path(loc, 'done - 2018-07-21.txt'))
tl <- bind_rows(tasks, done, done_old)
```

# Tasks completed per week

```{r completed-per-week}
dat <- tl %>%
  filter(done == 'x ') %>%
  mutate(week = as.numeric(format(date_completed, '%V')),
         year = as.numeric(format(date_completed, '%Y'))) %>%
  filter(!is.na(done)) %>%
  count(year, week) %>%
  filter(!is.na(year) | !is.na(week))

dat %>%
  complete_years() %>%
  complete(year, week) %>%
  ggplot(mapping = aes(x = week, y = n)) +
    geom_line(aes(group = 1), linewidth = 1.0, na.rm = FALSE) +
    facet_wrap(~ year) +
    labs(x = '\nWeek number', y = 'Number of task completed\n',
         title = 'Number of tasks completed per week\n') +
    theme_lemmens
```

Discovered that I've seem to lost about two and a half years worth of old tasks.

```{r completed-per-week-context}
n_colors <- tl %>% group_by_context() %>% distinct(context) %>% nrow()
long_palette <- colorRampPalette(RColorBrewer::brewer.pal(9, "Set1"))(n_colors)
tl %>%
  mutate(week = as.numeric(format(date_completed, '%V')),
         year = as.numeric(format(date_completed, '%Y'))) %>%
  filter(!is.na(done)) %>%
  group_by_context() %>%
    count(context, year, week) %>%
  ggplot(mapping = aes(x = week, y = n, colour = context)) + #, linetype = project)) +
    geom_line(size = 1.0, na.rm = TRUE) +
    facet_wrap(~ year) +
    labs(x = 'Week number', y = 'Number of task completed',
         title = 'Number of tasks completed per week\n') +
    scale_colour_manual(values = long_palette) +
    guides(colour = guide_legend(nrow = 2)) +
    theme_lemmens
```

# Creation / due

```{r create-due-lag}
create_due <- tl %>%
  filter(!is.na(date_created) & !is.na(date_due)) %>%
  mutate(create_lag = date_due - date_created)
create_due %>%
  group_by_context() %>%
  ggplot(mapping = aes(x = create_lag, fill = context)) +
    geom_histogram(binwidth = 1) +
    scale_fill_manual(values = long_palette) +
    coord_cartesian(xlim = c(0, 50)) +
    theme_lemmens
```


# Timeliness of completion

Not too large of a distance between actual completing a task versus when it should have been completed; i.e.
relatively few delays/lags.

```{r due-done-lag}
due_done <- tl %>%
  filter(!is.na(done)) %>%
  mutate(due_lag = date_completed - date_due)
p1 <- due_done %>%
  ggplot(mapping = aes(x = due_lag)) +
    geom_histogram(binwidth = 1) +
    theme_lemmens
  ## group_by(lst, tag) %>% use_series(due.lag)  %>% mean_cl_boot(due.lag)
p2 <- p1 %+% (due_done %>% filter(due_lag >= 0)) +
  coord_cartesian(xlim = c(0, 50))
p3 <- p1 %+% (due_done %>% filter(due_lag >= 1)) +
  coord_cartesian(xlim = c(0, 50))
p1 + p2 + p3
```

```{r due-done-by-context}
due_done %>%
  group_by_context() %>%
  filter(due_lag >= 1) %>%
  ggplot(mapping = aes(x = due_lag, fill = context)) +
    geom_histogram(binwidth = 1) +
    scale_fill_manual(values = long_palette) +
    coord_cartesian(xlim = c(0, 50)) +
    theme_lemmens
```



# Topics

```{r topics-over-time, fig.height = 9}
tl %>%
  mutate(week = format(date_created, '%V'),
         year = format(date_created, '%Y')) %>%
  mutate(yw = as.numeric(paste0(year, week))) %>%
  group_by_context() %>%
    group_by_project() %>%
      filter(!is.na(context) | !is.na(project)) %>%
      ggplot(mapping = aes(x = yw, y = project, colour = context)) +
        geom_line() +
        labs(x = '\nYearWeek#', y = 'Project tag\n') +
        scale_colour_manual(values = long_palette) +
        guides(colour = guide_legend(nrow = 2)) +
        theme_lemmens +
        theme(axis.text.y = element_text(size = rel(0.5)))
```