Provides functions and classes to compute, handle and visualise incidence from dated events for a defined time interval. Dates can be provided in various standard formats. The class 'incidence' is used to store computed incidence and can be easily manipulated, subsetted, and plotted. In addition, log-linear models can be fitted to 'incidence' objects using 'fit'. This package is part of the RECON (< https://www.repidemicsconsortium.org/>) toolkit for outbreak analysis.
To install the current stable, CRAN version of the package, type:
install.packages("incidence")
To benefit from the latest features and bug fixes, install the development, github version of the package using:
devtools::install_github("reconhub/incidence")
Note that this requires the package devtools installed.
The main features of the package include:
incidence
: compute incidence from dates in various formats; any fixed time interval can be used; the returned object is an instance of the (S3) class incidence.
plot
: this method (see ?plot.incidence
for details) plots incidence objects, and can also add predictions of the model(s) contained in an incidence_fit object (or a list of such objects).
fit
: fit one or two exponential models (i.e. linear regression on log-incidence) to an incidence object; two models are calibrated only if a date is provided to split the time series in two (argument split
); this is typically useful to model the two phases of exponential growth, and decrease of an outbreak; each model returned is an instance of the (S3) class incidence_fit, each of which contains various useful information (e.g. growth rate r, doubling/halving time, predictions and confidence intervals).
fit_optim_split
: finds the optimal date to split the time series in two, typically around the peak of the epidemic.
[
: lower-level subsetan of incidence objects, permiting to specify which dates and groups to retain; uses a syntax similar to matrices, i.e. x[i, j]
, where x
is the incidence object, i
a subset of dates, and j
a subset of groups.
subset
: subset an incidence object by specifying a time window.
pool
: pool incidence from different groups into one global incidence time series.
as.data.frame
: converts an incidence object into a data.frame
containing dates and incidence values.
An overview of incidence is provided below in the worked example below. More detailed tutorials are distributed as vignettes with the package:
vignette(package = "incidence")#> Vignettes not found
To open these, type:
vignette("overview", package="incidence")vignette("customize_plot", package="incidence")vignette("incidence_class", package="incidence")
The following websites are available:
The official incidence website, providing an overview of the package's functionalities, up-to-date tutorials and documentation:
http://www.repidemicsconsortium.org/incidence/
The incidence project on github, useful for developers, contributors, and users wanting to post issues, bug reports and feature requests:
http://github.com/reconhub/incidence
The incidence page on CRAN:
https://CRAN.R-project.org/package=incidence
Bug reports and feature requests should be posted on github using the issue system. All other questions should be posted on the RECON forum:
http://www.repidemicsconsortium.org/forum/
The following worked example provides a brief overview of the package's functionalities. See the vignettes section for more detailed tutorials.
This example uses the simulated Ebola Virus Disease (EVD) outbreak from the package outbreaks. We will compute incidence for various time steps, calibrate two exponential models around the peak of the epidemic, and analyse the results.
First, we load the data:
library(outbreaks)library(ggplot2)library(incidence) dat <- ebola.sim$linelist$date.of.onsetclass(dat)#> [1] "Date"head(dat)#> [1] "2014-04-07" "2014-04-15" "2014-04-21" "2014-04-27" "2014-04-26"#> [6] "2014-04-25"
We compute the weekly incidence:
i.7 <- incidence(dat, interval = 7)i.7#> <incidence object>#> [5888 cases from days 2014-04-07 to 2015-04-27]#> [5888 cases from ISO weeks 2014-W15 to 2015-W18]#> #> $counts: matrix with 56 rows and 1 columns#> $n: 5888 cases in total#> $dates: 56 dates marking the left-side of bins#> $interval: 7 days#> $timespan: 386 daysplot(i.7)
incidence
can also compute incidence by specified groups using the groups
argument. For instance, we can compute the weekly incidence by gender:
i.7.sex <- incidence(dat, interval = 7, groups = ebola.sim$linelist$gender)i.7.sex#> <incidence object>#> [5888 cases from days 2014-04-07 to 2015-04-27]#> [5888 cases from ISO weeks 2014-W15 to 2015-W18]#> [2 groups: f, m]#> #> $counts: matrix with 56 rows and 2 columns#> $n: 5888 cases in total#> $dates: 56 dates marking the left-side of bins#> $interval: 7 days#> $timespan: 386 daysplot(i.7.sex, stack = TRUE, border = "grey")
incidence
objectsincidence
objects can be manipulated easily. The [
operator implements subetting of dates (first argument) and groups (second argument).
For instance, to keep only the first 20 weeks of the epidemic:
i.7[1:20]#> <incidence object>#> [797 cases from days 2014-04-07 to 2014-08-18]#> [797 cases from ISO weeks 2014-W15 to 2014-W34]#> #> $counts: matrix with 20 rows and 1 columns#> $n: 797 cases in total#> $dates: 20 dates marking the left-side of bins#> $interval: 7 days#> $timespan: 134 daysplot(i.7[1:20])
Some temporal subsetting can be even simpler using subset
, which permits to retain data within a specified time window:
i.tail <- subset(i.7, from = as.Date("2015-01-01"))i.tail#> <incidence object>#> [1156 cases from days 2015-01-05 to 2015-04-27]#> [1156 cases from ISO weeks 2015-W02 to 2015-W18]#> #> $counts: matrix with 17 rows and 1 columns#> $n: 1156 cases in total#> $dates: 17 dates marking the left-side of bins#> $interval: 7 days#> $timespan: 113 daysplot(i.tail, border = "white")
Subsetting groups can also matter. For instance, let's try and visualise the incidence based on onset of symptoms by outcome:
i.7.outcome <- incidence(dat, 7, groups = ebola.sim$linelist$outcome)i.7.outcome#> <incidence object>#> [5888 cases from days 2014-04-07 to 2015-04-27]#> [5888 cases from ISO weeks 2014-W15 to 2015-W18]#> [3 groups: Death, NA, Recover]#> #> $counts: matrix with 56 rows and 3 columns#> $n: 5888 cases in total#> $dates: 56 dates marking the left-side of bins#> $interval: 7 days#> $timespan: 386 daysplot(i.7.outcome, stack = TRUE, border = "grey")
Groups can also be collapsed into a single time series using pool
:
i.pooled <- pool(i.7.outcome)i.pooled#> <incidence object>#> [5888 cases from days 2014-04-07 to 2015-04-27]#> [5888 cases from ISO weeks 2014-W15 to 2015-W18]#> #> $counts: matrix with 56 rows and 1 columns#> $n: 5888 cases in total#> $dates: 56 dates marking the left-side of bins#> $interval: 7 days#> $timespan: 386 daysidentical(i.7$counts, i.pooled$counts)#> [1] TRUE
Incidence data, excluding zeros, can be modelled using log-linear regression of the form: log(y) = r x t + b
where y is the incidence, r is the growth rate, t is the number of days since a specific point in time (typically the start of the outbreak), and b is the intercept.
Such model can be fitted to any incidence object using fit
.
Of course, a single log-linear model is not sufficient for modelling our time series, as there is clearly an growing and a decreasing phase.
As a start, we can calibrate a model on the first 20 weeks of the epidemic:
plot(i.7[1:20])
early.fit <- fit(i.7[1:20])early.fit#> <incidence_fit object>#> #> $lm: regression of log-incidence over time#> #> $info: list containing the following items:#> $r (daily growth rate):#> [1] 0.03175771#> #> $r.conf (confidence interval):#> 2.5 % 97.5 %#> [1,] 0.02596229 0.03755314#> #> $doubling (doubling time in days):#> [1] 21.8261#> #> $doubling.conf (confidence interval):#> 2.5 % 97.5 %#> [1,] 18.45777 26.69823#> #> $pred: data.frame of incidence predictions (20 rows, 5 columns)
The resulting objects can be plotted, in which case the prediction and its confidence interval is displayed:
plot(early.fit)
However, a better way to display these predictions is adding them to the incidence plot using the argument fit
:
plot(i.7[1:20], fit = early.fit)
In this case, we would ideally like to fit two models, before and after the peak of the epidemic. This is possible using the following approach, in which the best possible splitting date (i.e. the one maximizing the average fit of both models), is determined automatically:
best.fit <- fit_optim_split(i.7)best.fit#> $df#> dates mean.R2#> 1 2014-08-04 0.7650406#> 2 2014-08-11 0.8203351#> 3 2014-08-18 0.8598316#> 4 2014-08-25 0.8882682#> 5 2014-09-01 0.9120857#> 6 2014-09-08 0.9246023#> 7 2014-09-15 0.9338797#> 8 2014-09-22 0.9339813#> 9 2014-09-29 0.9333246#> 10 2014-10-06 0.9291131#> 11 2014-10-13 0.9232523#> 12 2014-10-20 0.9160439#> 13 2014-10-27 0.9071665#> #> $split#> [1] "2014-09-22"#> #> $fit#> $fit$before#> <incidence_fit object>#> #> $lm: regression of log-incidence over time#> #> $info: list containing the following items:#> $r (daily growth rate):#> [1] 0.02982209#> #> $r.conf (confidence interval):#> 2.5 % 97.5 %#> [1,] 0.02608945 0.03355474#> #> $doubling (doubling time in days):#> [1] 23.24274#> #> $doubling.conf (confidence interval):#> 2.5 % 97.5 %#> [1,] 20.65721 26.5681#> #> $pred: data.frame of incidence predictions (25 rows, 5 columns)#> #> $fit$after#> <incidence_fit object>#> #> $lm: regression of log-incidence over time#> #> $info: list containing the following items:#> $r (daily growth rate):#> [1] -0.01016191#> #> $r.conf (confidence interval):#> 2.5 % 97.5 %#> [1,] -0.01102526 -0.009298561#> #> $halving (halving time in days):#> [1] 68.21031#> #> $halving.conf (confidence interval):#> 2.5 % 97.5 %#> [1,] 62.86899 74.54349#> #> $pred: data.frame of incidence predictions (32 rows, 5 columns)#> #> #> $plot
plot(i.7, fit = best.fit$fit)
See details of contributions on:
https://github.com/reconhub/incidence/graphs/contributors
Contributions are welcome via pull requests.
Please note that this project is released with a Contributor Code of Conduct. By participating in this project you agree to abide by its terms.
Maintainer: Thibaut Jombart ([email protected])
incidence()
now returns an error when supplied a character vector that is
not formatted as (yyyy-mm-dd).
(See https://github.com/reconhub/incidence/issues/88)fit()
now returns correct coefficients when dates is POSIXt by converting to
Date. (See https://github.com/reconhub/incidence/issues/91)plot.incidence()
now plots in UTC by default for POSIXt incidence objects.
this prevents a bug where different time zones would cause a shift in the bars
(See https://github.com/reconhub/incidence/issues/99).first_date
and last_date
parameters.plot.incidence()
will now respect single groups.
(See https://github.com/reconhub/incidence/issues/84)as.data.frame.incidence()
will now respect single groups.
(See https://github.com/reconhub/incidence/issues/84)demo("incidence-demo" package = "incidence")
has been updated to show use of
custom colors.print.incidence()
will now print isoweeks even if the $interval
element is
"week".subset.incidence()
will now give a more informative error message when the
user specifies a group that does not exist.demo('incidence-demo', package = 'incidence')
now shows plotting with
show_cases = TRUE
.incidence.max.days
optionplot.incidence()
, incidence()
,
and as.data.frame.incidence()
For details, see
https://github.com/reconhub/incidence/issues/79group_names()
allows the user to retrieve and set the group names.get_timespan()
returns the $timespan
element.get_n()
returns the $n
element.dim()
, nrow()
, and ncol()
are now available for incidence objects,
returning the dimensions of the number of bins and the number of groups.plot()
called show_cases
has been added to draw borders
around individual cases for EPIET-style curves.
See https://github.com/reconhub/incidence/pull/72 for details.estimate_peak()
no longer fails with integer datesincidence()
no longer fails when providing both group information and a
first_date
or last_date
parameter that is inside the bounds of the
observed dates. Thanks to @mfaber for reporting this bug.
See https://github.com/reconhub/incidence/issues/70 for details.internal_checks.R
file has been split into the relative components.$lm
field of the incidence_fit
class is now named $model
to clearly
indicate that this can contain any model.incidence()
will now accept text-based intervals that are valid date
intervals: day, week, month, quarter, and year.
incidence()
now verifies that all user-supplied arguments are accurate
and spelled correctly.
fit_optim_split()
now gains a separate_split
argument that will determine
the optimal split separately for groups.
A new class, incidence_fit_list
, has been implemented to store and summarise
incidence_fit
objects within a nested list. This is the class returned by
in the $fit
element of fit_optim_split()
.
bootstrap()
will bootstrap epicurves stored as incidence
objects.
find_peak()
identifies the peak date of an incidence
objects.
estimate_peak()
uses bootstrap to estimate the peak time of a
partially observed outbreak.
get_interval()
will return the numeric interval or several
intervals in the case of intervals that can't be represented in a fixed
number of days (e.g. months).
get_dates()
returns the dates or counts of days on the right,
center, or left of the interval.
get_counts()
returns the matrix of case counts for each date.
get_fit()
returns a list of incidence_fit
objects from an
incidence_fit_list
object.
get_info()
returns information stored in the $info
element of an
incidence_fit
/incidence_fit_list
object.
incidence_fit_class
instructs the user on how
incidence_fit
and incidence_fit_list
objects are created and accessed.incidence()
function, the iso_week
parameter is deprecated in
favor of standard
for a more general way of indicating that the
interval should start at the beginning of a valid date timeframe.The $timespan
item in the incidence object from Dates was not type-stable
and would change if subsetted. A re-working of the incidence constructor
fixed this issue.
Misspelled or unrecgonized parameters passed to incidence()
will now cause
an error instead of being silently ignored.
Plotting for POSIXct data has been fixed.
incidence
object to avoid conflicts with additional
geoms such as geom_ribbon
, now used in projections::add_projections
.ggplot2
the argument n_breaks
has been added to plot.incidence
, to specify the
ideal number of breaks for the date legends; will work with ggplot2 > 2.2.1
added the internal function make_iso_weeks_breaks
to generate dates and
labels for date x-axis legends using ISO weeks
added a function add_incidence_fit
, which can be used for adding fits to
epicurves in a piping-friendly way
added a function cumulate
, which computes cumulative incidence and returns
an incidence
object
new generic as.incidence, to create incidence objects from already computed incidences. Methods for: matrix, data.frame, numeric vectors
better processing of input dates, including: automatic conversion from characters, issuing errors for factors, and silently converting numeric vectors which are essentially integers (issuing a warning otherwise)
new vignette on conversions to and from incidence objects
new tests
fixed issues caused by variables which changed names in some datasets of the outbreaks package, used in the documentation
disabled by default the isoweeks in incidence
; this part of the code will
break with changes made in the devel version of ggplot2, which is now
required by plotly
it is now possible to subset an incidence object based on Date
dates using
numeric values, which are interpreted as number of intervals since the first
date (origin = 1)
NAs are no longer removed from the input dates, as it would cause mismatches with grouping factors.
ebola.sim
->
ebola_sim
and ebola.sim.clean
-> ebola_sim_clean
add an argument iso_week
to incidence.Date() and incidence.POSIXt() to
support ISO week-based incidence when computing weekly incidence.
add an argument labels_iso_week
to plot.incidence() to label x axis tick
marks with ISO weeks when plotting ISO week-based weekly incidence.
The README.Rmd / README.md now contains information about various websites for incidence as well as guidelines for posting questions on the RECON forum.
incidence now has a dedicated website http://www.repidemicsconsortium.org/incidence/ generated with pkgdown
First release of the incidence package on CRAN!