---
title: "Plotting Minute-Level Activity Data"
output: rmarkdown::html_vignette
vignette: >
  %\VignetteIndexEntry{Plotting Minute-Level Activity Data}
  %\VignetteEngine{knitr::rmarkdown}
  %\VignetteEncoding{UTF-8}
---

```{r, include = FALSE}
knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.width = 14,
  fig.height = 7,
  out.width = "100%"
)
```

```{r setup}
library(actiplot)
```

This vignette uses the packaged `acti_minute_data` accelerometer recording.
It contains seven days of minute-level vector-magnitude activity counts.
It is processed from the [upper-limb activity Figshare
collection](https://doi.org/10.6084/m9.figshare.c.4457855)
described by [Chadwell et al. (2019)](https://doi.org/10.1038/s41597-019-0211-6).

```{r data}
activity <- acti_minute_data
```

## Complete recording

Use `acti_plot_time()` to inspect activity across the entire recording. The
`breaks` argument accepts ordinary duration strings and sets the interval of
the time-axis labels.

```{r full-time}
acti_plot_time(activity, counts, breaks = "4 hours")
```

For a more legible complete-recording axis, set `x_axis = "12 hours"`. It
shows date and 12-hour-clock labels at midnight and noon. Alternatively,
`x_axis = "midnight"` prints just the date at midnight, without `00:00:00`.

```{r full-time-readable-axis}
acti_plot_time(activity, counts, x_axis = "12 hours")
acti_plot_time(activity, counts, x_axis = "midnight")
```

## Daily aligned rows

Use `acti_plot_day()` to align each day on the same time-of-day axis. Each
facet row is one calendar date, which makes day-to-day activity patterns easy
to compare.

```{r daily, fig.height=12}
acti_plot_day(activity, counts, breaks = "4 hours")
```

Set `facet = "month-day"` to remove the year from date strips, or
`facet = "day"` to use the baseline-relative day calculated by
`actibase::acti_separate_times()`.

```{r daily-facet-labels, fig.height=12}
acti_plot_day(activity, counts, facet = "month-day")
acti_plot_day(activity, counts, facet = "day")
```

## Date-by-time heat map

`acti_plot_heatmap()` uses the same alignment with counts represented by fill.

```{r heatmap, fig.height=9}
acti_plot_heatmap(activity, counts, breaks = "4 hours")
```

## Actogram and daily profile

```{r actogram, fig.height=9}
acti_plot_actogram(activity, counts, breaks = "4 hours")
```

```{r minute-profile}
acti_plot_minute_profile(activity, counts, summary = "median")
```

## Interval annotations

Episode layers use an explicit interval table with POSIXct start and end
columns. The example creates deterministic nightly windows for illustration.

```{r intervals}
sleep_windows <- data.frame(
  start = as.POSIXct(c("2017-06-03 22:00:00", "2017-06-04 22:00:00"), tz = "GMT"),
  end = as.POSIXct(c("2017-06-04 07:00:00", "2017-06-05 07:00:00"), tz = "GMT")
)

acti_plot_time(activity, counts) +
  geom_sleep_window(sleep_windows) +
  geom_day_boundary("noon", linetype = "dashed")
```

The same interval schema supports generic episodes and wear annotations.

```{r episode-and-wear}
acti_plot_time(activity, counts) +
  geom_episode(sleep_windows, fill = "grey30", alpha = 0.10) +
  geom_wear(sleep_windows)
```

Additional lightweight annotations can mark selected activity and intensity.

```{r rug-threshold-vline}
acti_plot_time(activity, counts) +
  geom_activity_rug(activity[activity$counts > 10000, ]) +
  geom_threshold_band(activity, "counts", threshold = 10000) +
  geom_time_vline(xintercept = "12:00:00", linetype = "dotted")
```