Sometimes "data" is plural and sometimes it is singular - can we norm on one?

I saw that some chapters refer to "data science in education" (Chapter 2) and some chapters refer to "educational data science" (Chapter 4: Unique Challenges). Are they the same thing? If yes, should we norm on one term? If no, should we have definitions for them early on?

This was an idea we discussed during our call on August 7, 2018. The idea was to create a style guide to make our code consistent. We should also use the style guide to select one package when there are multiple packages that have similar purposes.

in compiling for bookdown the following error is showing up for chunks on lines:

• 171-175
• 185-195
• 201-211
• 278-285
• 320-332

Error: Required package is missing

I've set the code chunk to error = TRUE in order to facilitate rendering the text, but these will need to be addressed.

I'd like to draft a walkthrough using some data from students in an online science class.

The data, which I collected with a colleague, is anonymized--or at least I think it is, though I'd like feedback on this; see this PR, and is from around 650 students from around 20 science classes in a large, state-wide, public high school. The high school provides individual courses to students who cannot otherwise take them; it's not a full high school, if that makes sense. It includes "pre" and "post" survey data on students interest in science, value of science, and their perceptions of their self-confidence/competence. It also includes a few "trace" measures, namely, how much time they spent in the LMS.

I would like to use this for one of the three planned walkthroughs and possibly for the section of the advanced uses chapter on multi-level or mixed effects models

This chapter is an odd duck: every other chapter focuses upon teachers/administrators/researchers/policymakers using data science techniques (and R), whereas this focuses more upon teaching individuals at the K-12 age.

I wonder if in some ways we can suggest that some of the ideas and techniques used in this book could be used to teach youth/students, i.e. some of the ideas potential ideas for chapter 10 discussed in #36. Thoughts?

In addition to possibly recapitulating (and adapting for younger R learners), some possible distinctions that may make this chapter different could be:

• K-12 students often need to meet standards
• There aren't stand-alone data science standards (though there are many elements in the science, math, and even computer science standards, i.e. modeling and information literacy)

During yesterday's call we noticed the narrative structure of the book coming together. We talked about how groups of chapters could serve the main points. These are the basic ideas:

Part 1: Chapters 1-4

• Educators who use data science techniques face unique challenges
• It takes a lot of people in different roles to create a good learning environment. Different education roles mean different datasets, techniques, and challenges

Part 2: Chapters 5-9

• Readers can learn and practice steps in the analysis process
• The walk-throughs use education datasets that feel familiar to education workers. Readers can work through examples that feel more practical than commonly used datasets like mtcars or iris
• Resources like R for Data Science, blog posts, and data science forums are part of the ongoing learning process

A note on the tone of this section: This section can serve the same purpose as blog posts like this one. Readers can get practical instruction, then deepen their learning by reading more in-depth materials.

Part 3: Chapters 10-12

• Now that educators are familiar with some data science techniques, it's time to tackle applying and sharing those techniques at work. This is its own skill.
• Connect the skills taught in Part 2 to the unique challenges discussed in Part 1
• Explore how to use programming, statistics, and content knowledge in an education workplace that is new to data science
• Explore how educators can teach today's students to prepare them for using data in future work places

We can use this issue thread to discuss and develop these ideas. I'll follow the conversation here and will eventually move the narrative structure into the readme. Then writers can use the structure to focus their work on the main point of a chapter. They can also zoom out and see how well the chapter contributes to its section of the book.

Feel free to discuss!

Need to do this after #114 gets merged

From Slack:

wonder if we could explicitly discuss who works with data and how they do so (maybe in the intro.), i.e. what types of data do administrators OR data analysts OR researchers primarily work with. And maybe we can suggest specific chapters that will be particularly relevant. Part of this could be delimiting who works with data but who is not a focus of our audience (i.e., parents, and, to a lesser extent cuz we touch on this, teachers and students)

Maybe this can be in chapter 2 under “pick a chapter that’s useful for you”?

New content for walkthrough 5, which uses publicly available NAEP data.

Use @jmostipak 's spreadsheet to catalogue. Then we can coordinate initial explanations and future references.

Walkthrough 1 (06-walkthrough-1.Rmd) has a foundation, but has some room for improvement, especially with respect to the newer parts on visualizations and models.

Right now, these focus on a fairly obvious relationship between the time students spent on the course and the percentage of possible points earned. There are other relationships that could be more interesting. A few ideas:

• Look at differences by the five subjects (anatomy and physiology, physics, biology, chemistry, and forensic science)
  • To do this analysis, the code for the subject would have to be identified from the course-ID; this would take a little regex work
• Look at some of the relations of the self-report variables for interest, value, and self-efficacy (perhaps with time spent on course)

We could also look at gender, though it's really sex, and I'm concerned about treating this variable as a token dichotomous/dummy code one in any case (but am open to feedback on this!).

• Review the file for chapter 8 (08-walkthrough-3.Rmd) to see which variables are included - in any way at all - in the analysis
• Record what those variables are
• Subset the online science motivation dataset, used in that walkthrough and stored in data/online-science-motivation/processed/online-science-motivation.csv, to include only those variables used in the analysis
• Use synthpop::syn() on the subset data frame to synthesize a new data frame
• Replace the online science motivation dataset with the new, synthesized dataset

The RMarkdown file gets stuck when executing this

source("r/theme_dataedu.R")
source("r/palette_dataedu.R")

I see this in the console and I type y
Importing fonts may take a few minutes, depending on the number of fonts and the speed of the system. Continue? [y/n] y

Then it returns this:

input string 4 is invalid in this localeinput string 4 is invalid in this localeinput string 4 is invalid in this localeinput string 4 is invalid in this locale

From my friend @markubsch via Twitter:

I recently looked at the GitHub page for your data science book. Have you though of having a chapter on bayesian methods as well? I think there are some arguments that make Bayes especially interesting in education.

New content for walkthrough 2, which uses gradebook data.

happy to take on outlining and beginning work on this chapter!

Hi all! I've started making plots on a chapter. Would we be interested in having a style guide for ggplots? Happy to lead on creating one!

At some point, we need to connect datascienceineducation.com with our Netlify-hosted gitbook. I think this would involve changing the nameservers to point to netlify - and then changing some Netlify setting to use our custom URL.

Seperately, would it make any sense to also purchase dsieur.com?

Hi all, curious what thoughts you all have on potential data sets to use in examples. I added a file to the planning folder with just a few bare-bones ideas (see here), but we can edit that in conjunction with discussing this here or elsewhere.

in reading through walkthrough 6 I noticed that the code switched back and forth between base R and the tidyverse styles in cases where tidyverse solutions are available. depending on the intended audience for the text it may be beneficial to choose a single syntax structure and use it consistently throughout the text.

I'm happy to take a pass at some of the code formatting if that would be helpful!

Hi again,

I can most contribute on topics regarding wrangling public datasets. I've worked with NAEP and would be glad to take on that chapter. Another thing I was thinking was showing imputation with EdFacts graduation data, which is suppressed when there are small cell sizes. Let me know if this is applicable to your work and how I can get started!

Hi all, was thinking of outlining / drafting this chapter (chapter 4). If it's okay for me to work on, will share more thoughts and ideas as I begin to make progress.

It might be useful to include a section on EDA within the context of gradebooks. I suspect thousands of teachers and department chairs have exported a gradebook, opened the spreadsheet, and wondered what to do next. I myself have wondered what to look for when combing through summative and formative assessments.

Please ignore the vulgarity of the code, but I ventured a small tutorial for the Michigan Virtual Learning Research Institute (MVLRI) here. I would be interested to read a smarter person's take on something like this.

I think the path needs to be corrected to the dataset on this one. Once all the code runs we can change the file name and get it rendered in bookdown

Can you please send me a DM on Slack (RyanEs) with a short description about the data science work you do in education?

I'm working on chapter 3, which is about the role of a data scientist in education. I want to get some examples of what y'all do in your day-to-day work and what your responsibilities are as someone who does data science in the education field. I think sharing different examples would tell the story about data science in education better than trying to nail it down to single definition.

After I collect a few, I'll write them up then try and draw some conclusions from the collection of activities.

Just a few ideas for this chapter--open to (and requiring) more ideas & input:

• Have a project/problem to work on
• Start where people are (with the tools they use and the data they have - as well as the problems they are trying to solve with data science/R)
• Recognize that using data science tools such as R can take a long time and an investment of effort

Hi all, I have an opportunity to discuss 'open science' as part of a proposal to present at the American Educational Research Association (AERA) conference.

My first thought was to discuss some of my experiences writing this book. I've found it rewarding and perspective-broadening to write this collaboratively and in the open. Though it's not finished (and is very much still in-progress), given the chance, I'd like to share what we've done, but I'd like your thumbs up to do so. If you have thoughts (positive, negative, or other!) about this, please let me know!

Also, if folks would like to join in the presentation (helping with the proposal and, if you like and if it is accepted, presenting at the conference in Toronto in April), I would welcome that (it's not necessary to be at the conference to be involved and at least for AERA many authors aren't able to attend) .

This book is awesome. Found a typo in walkthrough 1: in the code chunk on lines 205-213:

# split course section into components
s12_course_data <- s12_course_data %>%
  separate(col = CourseSectionOrigId,
           into = c('subject', 'semester', 'section'),
           sep = '-',
           remove = FALSE)

CourseSectionOrigId should be CourseSectionOrigID, otherwise it's not found.

Does anyone else find it sort of strange that Advanced Uses (06-advanced-uses.Rmd) comes before our walkthroughs? Can we move this--or integrate the content that would be in this into the walkthroughs?

For walkthrough 4 (09-walkthrough-4.Rmd), there is a problem with the simulated data. The key when generating the data is that nominators with higher values of yvar2 need to have relations with - have nominated - nominees with higher levels of yvar1. This isn't the case yet.

Thank you @ivelasq, @kierisi, and @restrellado for your thoughts on #28. Here are the things that based on the feedback that I think need to be done:

1. The idea of having two tracks seems to have some traction: I like this and forgot that some our readers will have experience using R - or at least may have used it a few times, attended a workshop, or been involved with a collaborator who used it. I think the following sounds great:
   • one can be 'installation and setup of R and R Studio', which covers installation, projects, and packages
   • the second can be 'data loading and manipulation using the tidyverse', which covers reading/saving files, pipes, selecting, filtering, etc. chapters.
2. I concur that there's a lot of ways to load different data but not a descriptive of what to use when; I think we can and should address this.
3. Finally, I think we should mention more about what purpose swirl or DataCamp may serve to our audience.

I'll leave this issue open for a little while (is one week okay?) and then will plan to make these changes.

Should we add a conclusion chapter?

The online science classes data set (used in walkthrough 1 and walkthrough 3 is anonymous but is actual data I collected with a colleague (a post-doc) at MSU.

I think it would be preferable to simulate data based upon this data - instead of sharing it directly.

The code in walkthrough 1 needs to be run locally to see if everything works as written.

We need some instructions for using devtools::install_github("data-edu/dataedu")

I clicked on slack link to go into a login page. How do I get invited to the channel?

for now I've set the code chunk to ignore the error when compiling using

Quitting from lines 177-184 (06-walkthrough-1.Rmd) 
Error in eval_tidy(enquo(var), var_env) : 
  object 'CourseSectionOrigId' not found
Calls: local ... separate -> separate.data.frame -> <Anonymous> -> eval_tidy
In addition: Warning message:
funs() is soft deprecated as of dplyr 0.8.0
please use list() instead