Lazymodel

Description

The lazymodel package allows you to easily build models to permform statistical analysis without the need of writting a lots of R code.

This depends on three other packages to work:

• glmmTMB (>= 1.0.1)
• argparser (>= 0.7.1)
• DHARMa (>= 0.2.7)

And needs R >= 3.4.4 to work.

Installation

To install this package, the devtools package must be installed.

Run in R the following command to install the package:

> install_gitlab("nfontrod/lazymodel", host = "https://gitbio.ens-lyon.fr", quiet = FALSE)

Usage

In R

Here is a small example on how to use the package in R:

> library(lazymodel)
> student <- data("student", package="lazymodel") # fake dataset of 200 students with their size, age and sex.
> head(student)
      size age sex
1 191.6936  24 Man
2 205.1404  23 Man
3 175.1561  19 Man
4 161.7353  18 Man
5 186.2363  20 Man
6 181.8675  25 Man
> #?make_analysis: to display the help
> make_analysis("size ~ age + sex", family="gaussian", data=student, output="./") # Analysis to see if the age or the sex of students as an impact on their size

The previous function will produce a diagnotics figures that allows you to see if your model is valid. If you see in the right panel of this figure a non-significant deviation (KS test p>0.05), and red lines in the “residuals vs predicted” plots that match quite well the black dotted lines, this means the model is a good fit to the data.

If the model is a good fit, you can look at the p-values given in the summary file of your current directory.

With a CLI (command line interface) script

First, you must create an R file containing only the following code:

# my_R_file.R
library(lazymodel)
cli_analysis()

Then you can type the following commands to see if everything works:

Rscript my_R_file.R --help

This command should display the following content:

usage: test.R [--] [--help] [--opts OPTS] [--table TABLE] [--formula
       FORMULA] [--family FAMILY] [--output OUTPUT]

Model your data according to a given distribution and then performs a
statistical analysis.

flags:
  -h, --help     show this help message and exit

optional arguments:
  -x, --opts     RDS file containing argument values
  -t, --table    A file containing a table inside in which each column
                 is separated by tabs
  -f, --formula  The formula of your model
  -F, --family   The family name of the distribution tested [default:
                 gaussian]
  -o, --output   The folder that will contains the results of the
                 analysis [default: ./]

You have to pass to --table argument, a file containing a table (columns separated by tabulations) with column header. You have to pass to --formula argument, a string corresponding to an R formula (e.g response_variable ~ pred1 + pred2). Note that the name given in the formula must be the same as those in the column header of the file given with the --table argument.