Skip to content

svitkin/readme-example

Repository files navigation

About

A template repository for fully automated model documentation through a README file. Detailed blog post here.

Modeling code is in R/model.R. To enrich your README with model documentation, simply add a comment in the following format above the modeling call.

#~ A Section Header I want the model results to go under
m <- lm(mpg ~ cyl + hp, data = mtcars)

To group multiple models into one section for comparison, use the same section name and explicitly name the model in {} brackets.

#~ Modeling MPG {Without AM}
m <- lm(mpg ~ cyl + hp, data = mtcars)

#~ Modeling MPG {With AM}
m <- lm(mpg ~ cyl + hp + am, data = mtcars)

You can either source model_readme/process_file.R or push your change to automatically update the README. Code to automatically document models is located in the model_readme directory. Currently, for a given section, the following is included in the README:

  • Model summary values from glance
  • Coefficient estimates from tidy
  • Coefficient forest plot (OPTIONAL)
  • Residuals vs Fitted Values plot

All documentation below is automatically generated with this repository.

Models

MPG Model with mtcars

Summary

Using AM Without AM
r.squared 0.804 0.741
adj.r.squared 0.783 0.723
sigma 2.807 3.173
statistic 38.319 41.422
p.value 0.000 0.000
df 3.000 2.000
logLik -76.292 -80.781
AIC 162.585 169.562
BIC 169.914 175.425
deviance 220.553 291.975
df.residual 28.000 29.000
nobs 32.000 32.000

Components

Characteristic Using AM Without AM
Beta 95% CI1 p-value Beta 95% CI1 p-value
cyl -1.1 -2.4, 0.17 0.086 -2.3 -3.4, -1.1 <0.001
am 3.9 1.2, 6.6 0.005
hp -0.04 -0.07, -0.01 0.017 -0.02 -0.05, 0.01 0.2
1 CI = Confidence Interval

Residuals

Logit AM Model

Summary

null.deviance 43.230
df.null 31.000
logLik -9.173
AIC 26.347
BIC 32.210
deviance 18.347
df.residual 28.000
nobs 32.000

Components

Characteristic log(OR)1 95% CI1 p-value
cyl -0.63 -2.1, 0.68 0.4
mpg 1.1 0.28, 2.6 0.062
hp 0.06 0.02, 0.15 0.039
1 OR = Odds Ratio, CI = Confidence Interval

Residuals

AQ Model

Summary

r.squared 0.606
adj.r.squared 0.595
sigma 21.181
statistic 54.834
p.value 0.000
df 3.000
logLik -494.359
AIC 998.717
BIC 1012.265
deviance 48002.790
df.residual 107.000
nobs 111.000

Components

Characteristic Beta 95% CI1 p-value
Wind -3.3 -4.6, -2.0 <0.001
Solar.R 0.06 0.01, 0.11 0.011
Temp 1.7 1.1, 2.2 <0.001
1 CI = Confidence Interval

Residuals

About

An example repository for automatic model documentation in a README

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages