openintrostat / openintro Goto Github PK

Contact Details

[email protected]

Bug

The fastfood data set has a salad variable with all 515 values "Other". Looking at the item descriptions, it does appear that there are actual salads in the data set.

Reproducible Example

library(openintro)
#> Loading required package: airports
#> Loading required package: cherryblossom
#> Loading required package: usdata
table(fastfood$salad)
#>
#> Other
#> 515

Expected Behavior

I expected to see some foods classified as salads and others, not.

Session Info

No response

Additional context

No response

Hi,

Please I don't find bat10 dataset use in the ANOVA of chapter 5 Inference for numerical data

Where could I download it please ?

On page 191 the Fourth Edition of the textbook mentions the rosling_responses data set:

"We will use the rosling_responses data set to evaluate the hypothesis test ..."

Use of the texttt font for "rosling_responses" suggests that such a data set exists in the package, but it doesn't.

it says "something is wrong" but nothing is wrong

Hi,
the yrbss data is used in the OpenIntro text.
The yrbss data is available to download on the Github site.
So far as I can tell, the yrbss data hasn't been added to the OpenIntro packages.
Should it be?

Referring to the text that says "Please visit openintro.org for free statistics". It shows up in the compiled markdown documents (as shown below), and yes, it's possible to mute that with the message = FALSE option in the chunk, but I think we want to be careful about teaching those to students who are new to R.

Leaving the issue here to be consider before the next version of the package...

Nick has added a handful of data sets.

Hi,

When I render the image with Korean Character, the Korean characters are broken.
-. myPDF in variable.R

However, for instance, when I test the CairoPDF, the Korean characters are rendered correctly, but there are width and height issues.

I think that the other asian characters will have similar issues when using openintro package.

Thank you.

nuff said

From looking at your examples, I'm not exactly sure what the purpose of your dotPlot() is supposed to be, but it is unfortunate that you have chosen a name that conflicts with the version in the mosaic package, which makes the kind of dot plot often seen in introductory statistics courses.

mosaic::dotPlot( ~ rnorm(500), width = 0.1)

Revisit the format of the data and the examples provided

For example, rather than give probabilities, show how 1000 cases would cascade through the tree.

it's not documented

Unless I am missing it, this is neither the births nor the ncbirths data set.

Side note: Is it necessary to have both births and ncbirths?

Is there a place where I can find the R code by chapter for the openintro book ?

Make dataset one word in DESCRIPTION

Are observations people, days, something else? Docs will need to be updated accordingly.

Do we know which year's survey is included in this dataset? Also, do we know if the variable called gender is what's identified in the 2017 data documentation as sex?

I'm happy to do a PR to clarify those things if we can track them down.

@mine-cetinkaya-rundel can we push a new version of openintro to CRAN? We're using the latest version from github in your datacamp courses, but I think there's been some confusion among students since it differs from what's on CRAN.

This seems unnecessary and confusing:

library(openintro)
## Please visit openintro.org for free statistics materials
## 
## Attaching package: ‘openintro’
## 
## The following objects are masked from ‘package:datasets’:
## 
##     cars, chickwts, trees

This would be helpful for non-R users of the datasets.

@DavidDiez I know you host these on openintro.org but keeping synced seems a challenge. I could automate it here and post on the package websites and openintro.org could point to them. Or I suppose you could build the page on your end based on the automatically generated files in this repo as well. We should discuss which approach is preferable, but at least automatically generating files as we update the package seems like a good idea.

From Jack Miller:

The blood alcohol data set has been around since 1992 and appeared in the Electronic Encyclopedia of Statistical Examples and Exercises. I worked on EESEE and used the data sets at OSU, so I am very familiar with that particular citation. :-) Here is a URL for that particular "story" in EESEE: http://bcs.whfreeman.com/WebPub/Statistics/shared_resources/EESEE/BloodAlcoholContent/index.html.

This change will need to propagate to IMS and other books that reference this dataset as well.

Says absenteeism, should say Human Freedom Index

R Markdown has the ability to create templates. Could we create a lab report template for the On Your Own questions that instructors could use in their classes?

...and make it a dependency of OpenIntro

Contact Details

No response

Source

The aldrin dataset from slides is not in the package.

Source is at https://github.com/OpenIntroStat/openintro-statistics-slides/tree/master/Chp%207/7-5_anova/figures/aldrin.

Desired Solution

No response

Alternative Solutions

No response

Additional context

No response

• In both email and email50 there are variables in the docs that don't exist in the data: period_mess and signoff -- should be removed from docs
• email50 example code yields FALSE (random sampling change might be the cause?)
• In both datasets indicator variables should be factors
• cc is numeric, not indicator

Current in \dontrun{} but they need to be checked and rewritten

1. Use scales == "free" or better, add a scales argument that defaults to "free". [Else a sample with an outlier will cause the other plots to look quite different from how they would look if they were generated in isolation.]

2. Don't hard code the number of simulations. Let 8 be the default if you like.

3. rename first argument? It's a bit of an odd name. But I'm guessing it will typically be used without naming, so this is not such a big deal.

4. Consider a version that doesn't label the original data but makes it one of the sample (randomly selecting which location). Not sure the best way to do the "reveal".

5. Perhaps add a seed argument that sets the seed used. That would solve the reveal issue in one way, since the plot could be generated again withe the original data set distinguished.

6. Complete the documentation and include examples.