Width argument to specify no gap between bars about incidence2 HOT 7 CLOSED

Hello, good suggestion re: the width.

I'm pretty sure the irregularity of the white lines is being caused by the resolution of the saved image. If you recreate the plot in R and look at it interactively (and on a large screen) you will see the lines are evenly spaced. I will regenerate the vignettes to fix the issue.

In regards to the width argument, I will add the argument to plot.incidence2 and probably also to scale_x_incidence. It may be sensible to default to full width (I think ggplot2 defaults to 90%) as this would avoid any issue with resolution. This should be straightforward for yrwk, yrmon yrqtr but may be a little more fiddly for arbitrary periods.

For info, I've split off the logic for the grouped dates in to a separate package called grates. This package has zero strong dependencies so if you didn't want the overhead of incidence2 you may like that. It also fixes some bugs I found in the current master branch of incidence2 plus has expanded documentation on the class implementations. This week I'll be dropping the current implementation and pulling in grates as a dependency going forward.

We're closing in on the stable 1.0 release it's now just ironing out these little issues like widths etc. Hopefully within the next week or two assuming nothing urgent comes up.

from incidence2.

Thanks for the follow-up Tim, I look forward to the 1.0 release. And for sharing grates - good to know!

I've also realized that my original temporary solution of providing width = 7 to geom_col() won't work for months, as each month has a different number of days. But it sounds like there is more complexity going on in the background with grates etc, so I hope the width 100% solution is straight-forward for you.

Thanks again,

from incidence2.

yep - when I integrate grates this should all just work by setting width = 1 as a yrwk, yrqtr and yrmon will always be treated as 1 unit (see below). Using width = 1 should also work in the current master branch of incidence2. Where it doesn't work is for arbitrary periods of time (created with as_period) but I'm hoping I can hide the additional complexity for this situation from the users.

library(grates)
library(outbreaks)
library(data.table)
library(tibble)
library(ggplot2)

# load data
dat <- as.data.table(ebola_sim_clean$linelist)

# yearweek
yearweek <- dat[, isoweek := as_yrwk(date_of_onset)
                ][, .(cases_new = .N), keyby = isoweek]
as_tibble(yearweek)
#> # A tibble: 56 x 2
#>    isoweek  cases_new
#>    <yrwk>       <int>
#>  1 2014-W15         1
#>  2 2014-W16         1
#>  3 2014-W17         5
#>  4 2014-W18         4
#>  5 2014-W19        12
#>  6 2014-W20        17
#>  7 2014-W21        15
#>  8 2014-W22        19
#>  9 2014-W23        23
#> 10 2014-W24        21
#> # … with 46 more rows
ggplot(yearweek, aes(isoweek, cases_new)) + geom_col(width=1) + theme_bw() + xlab("")

# yearmonth
yearmonth <- dat[, yrmon := as_yrmon(date_of_onset)
                 ][, .(cases_new = .N), keyby = yrmon]
as_tibble(yearmonth)
#> # A tibble: 13 x 2
#>    yrmon    cases_new
#>    <yrmon>      <int>
#>  1 2014-Apr         7
#>  2 2014-May        67
#>  3 2014-Jun       102
#>  4 2014-Jul       228
#>  5 2014-Aug       540
#>  6 2014-Sep      1144
#>  7 2014-Oct      1199
#>  8 2014-Nov       779
#>  9 2014-Dec       567
#> 10 2015-Jan       427
#> 11 2015-Feb       307
#> 12 2015-Mar       277
#> 13 2015-Apr       185
ggplot(yearmonth, aes(yrmon, cases_new)) + geom_col(width=1) + theme_bw() + xlab("")

^{Created on 2021-03-21 by the reprex package (v1.0.0)}

from incidence2.

Perfect - thank you!

Perhaps I misunderstood your reply ("Using width = 1 should also work in the current master branch of incidence2."), but I just tried using width = 1 with the current master branch of incidence2 and the argument wasn't accepted. Did I do something wrong?

# install latest master branch
pacman::p_load_current_gh("reconhub/incidence2@master")

# load data
linelist <- as.data.table(ebola_sim_clean$linelist)

# make incidence
a <- incidence2::incidence(linelist, date_of_onset, interval = "week", groups = gender)

# plot
facet_plot(a)
plot(a, width = 1)  # did not work

Error was: Error in ggplot2::scale_x_continuous(..., trans = yrwk_trans(n = n, firstday = firstday)) : unused argument (width = 1)

Also, I tried to install grates but the repo was an "unknown package". I'm happy to wait till later to try your example above, but I just wanted you to know. Perhaps there are permissions? I used remotes::install_github("reconhub/grates") from https://www.reconverse.org/grates/.

from incidence2.

Regarding grates - I've realized that it's just that "reconhub" should be "reconverse" in the install command in the README.
I got it to install with remotes::install_github("reconverse/grates")

from incidence2.

This should hopefully be fixed (as best possible) by 3361373. To see the new functionality you will need the master branch of both grates and incidence2.

When the interval is numeric, "x days", "x weeks", "month", "quarter" or "year" and you pass the argument color=NA to plot this should now work as desired. Intervals such as "x months" (where x >= 2) are causing me trouble so I'll raise a separate issue to track this. Dense plots may still have the white line artefacts but this can hopefully be resolved by users by plotting higher resolution images.

Example:

library(outbreaks)
library(incidence2)

dat <- ebola_sim_clean$linelist

i_week_sex <- incidence(
  dat,
  interval = "week",
  date_index = date_of_onset,
  groups = gender
)

# no lines
facet_plot(i_week_sex, n_breaks = 3, color = NA)

# with lines
facet_plot(i_week_sex, n_breaks = 3)

Is this the sort of behaviour you were after?

from incidence2.

from incidence2.