In my last post, I showed how to extend the functions I developed for NJASK data to other state assessments. In this post, I'll tie eveything together, and write some general functions that bring a wide variety of NJ state assessment data into R.

Roughly speaking, we're tying to write a function that will return data given a year and a grade. Here are the big things that need to happen:

1. Check that call we made is a valid grade/year combination (raising an informative error if not)

2. Map the grade / year call to the correct get_blank_data function (NJASK? HSPA? GEPA?)

3. Fetch, clean, and return the data frame.

Let's tackle each of these in turn:.

In my last post, I talked about how to programmatically process and cleanup NJASK data. In this post, we'll extend the NJASK functions to the High School Proficiency Assessment (HSPA), and to the old Grade Eight Proficiency Assessment (GEPA). With functions that can access each of those data sources, we'll be ready to write a general wrapper that simplifies access to relevant state assessment data.

In my last post, I showed how to read a fixed width file into R. In this post, I'll refactor that code into a function that takes year/grade as a parameter.

First, read in the layout data frame. This contains the column headers for NJASK fixed with files.

New Jersey has switched over to the PARCC assessment, but from roughly 2005-2014 New Jersey took the "New Jersey Assessment of Skills and Knowledge," or more simply "the NJASK."

This data is a bit of a pain to work with. There's a page for each year of assessment data, and for each year there's a page per grade level. But wait! It gets better! If you download the Excel spreadsheet for each grade level (helpfully titled state_summary.xls for every year and grade!), you'll find 5 tabs per file, because there are simply too many demographic/subject subgroups for Excel 2003 to handle.

I leaned heavily on a great post from Rebecca Weiss titled "How easy is it to use R markdown and knitr with pelican?" to get Pelican/knitr integration up and running. Rebecca wrote her post in 2014 in a Mac/Unix environment, so I thought that I would contribute an update, and some errata for Windows users.

If you're landing here from Google, a quick overview & definition of some terms:

• Pelican is a static site generator written in Python. If you've ever seen the Github pages tutorials that use Jekyll, Pelican is a great alternative if you want to stay in the Python universe.

• knitr is a report generation package for R (especially R Studio) that makes it easy to mix words, simple formatting and R code.

It would be great to be able to write up an .Rmd, run pelican content, and have everything just show up on the web, with syntax highlighting, inline images, etc. That's what Pelican + knitr lets us do!

just a test post to see if pelican works…