bueller

Assessment Participation & Chronic Absenteeism Analytics for State Education Agencies

Things I Want You to Know

  1. What the basic idea behind the package is

  2. What kind of data the package needs

  3. What the package does

  4. What the output looks like

  5. What additional features I’d like to add

  6. What I’ve found challenging so far

My Question for You

What must states understand about their schools before addressing absenteeism and test participation?

The Basic Idea

SEA analyst has properly formatted data

↓

SEA analyst types bueller(data); <Enter>

↓

SEA analyst gets pretty reports & visuals

↓

SEA analyst has more time to do other important things

Example Data

  • Source: California Department of Education public data files
  • Years: 2018-19 through 2023-24 (Missing 2019-20)
  • Chronic Absenteeism Files:
    • Number of students eligible to be counted
    • Number of students chronically absent
    • Chronic absenteeism rate
  • Assessment Files:
    • Number of enrolled students
    • Number of tested students (ELA & Math)
    • Participation rate
  • School-level data: Each row is a school-year-group observation
  • Data available as far back as 2016-17

What the Package Does

# Install the package
devtools::install_github("erwx/bueller")

# Load the package
library(bueller)

# Generate your report
bueller(data = your_prepared_data)

  • Simple function call: Just point to the data

  • Helper function automatically sets up directory structure to store output files

What the Package Does

  • How many districts have had participation rates lower than 90% in each of the last 5 years?
  • Is the number of schools with chronic absenteeism rates greater than 30% increasing or decreasing?
  • Are a few schools driving the low participation rates in this district?
  • Which student populations are systematically under-represented in testing data?
  • Are engagement gaps widening or narrowing over time?

Pretty Reports

Additional Features I’d Like to Add

  • Like Schools/Like Districts: Group schools/districts to get more meaningful comparisons of absenteeism and participation rates
  • American Community Survey API: Zip-code level demographic and economic context
  • RevealJS presentations: Automated slide decks for stakeholder meetings
  • Going Mobile: Make a platform that reports relevant metrics directly to non-technical stakeholders

Challenges

  • SEA analysts already know what they care about
    • How do I decide what is “important”
  • Ease of Use vs. Flexibility
    • Everyone likes things that just work…
    • …but savvy users might want more options
  • R interfaces nicely with AI models
    • Do I board the AI hype train?