class: center, middle, inverse, title-slide .title[ # Consolidating Analysis and Dissemination of Research using GitHub ] .subtitle[ ## Using GitHub for Open-Source Analytics, Reporting, and Dissemination of Research ] .author[ ### Nathan Dadey ] .institute[ ### Center for Assessment ] .date[ ### April 23rd, 2022
NCME 2022 Annual Meeting
Demonstration Session
] --- layout: true --- class: inverse, center, middle # Consolidating Analysis and Dissemination of Research using GitHub --- # Motivation: Data Driven Documents + In many cases, work can be shared well via one or more **short vignettes or other write ups** (e.g., Allie's excellent vignettes on [multiple imputation](https://centerforassessment.github.io/Internship_2021/articles/PSW_Vignette_Landing_Page.html) and [propensity score weighting](https://centerforassessment.github.io/Internship_2021/articles/MI_Vignette_Landing_Page.html)) -- + Such write-ups can also be thought of as one form of **data-driven documents** (e.g., [NYT](https:/www.nytimes.com/interactive/2017/12/05/upshot/a-better-way-to-compare-public-schools.html) [Articles](https:/www.nytimes.com/interactive/2017/12/05/upshot/a-better-way-to-compare-public-schools.html), which typically: + Are well structured + Provide a limited n umber of comparisons + Meant to tell a predetermined story -- + This approach has been implemented in the in the prior presentations. --- # Motivation: Dashboards and Data Explorers + In other cases, we want to encourage users to **interact with data**, either through dashboards (e.g., [CO](https://edx.cde.state.co.us/SchoolView/DataCenter/reports.jspx), [CA](https://www.cde.ca.gov/ta/ac/cm/), [NYC](https://tools.nycenet.edu/), [MI](https://www.mischooldata.org/DistrictSchoolProfiles2/ReportCard/EducationDashboard4.aspx)) or more elaborate data exploration tools (e.g., [NAEP](https://www.nationsreportcard.gov/ndecore/landing)) -- + These kinds of tools typically: + Are flexible + Support numerous comparisons + Are DIY, allowing users to find their own stories -- + Implementing this approach may involve [flexdashboard](https://rmarkdown.rstudio.com/flexdashboard/) with interactivity supported through packages like [DT](https://rstudio.github.io/DT/), [polt.ly](https://plot.ly/r/) , [htmlwidgets](https://www.htmlwidgets.org/) or Shiny (e.g., see this [simple example](https://www.lexjansen.com/pharmasug-cn/2019/DV/Pharmasug-China-2019-DV19.pdf)). --- # Motivation: A Third Use Case + In this final case, the **full complexities of multiple analysis** may need to be captured (e.g., a larger set of connected write ups): + Well structured across numerous write ups + Contain multiple comparisons supported through multiple analyses + Meant to provide a balance between structured and DIY storytelling -- + One way to implement this approach is by binding up the multiple write ups from multiple analyses (captured in individual . rmds ) into a larger book via the [bookdown](https://bookdown.org/yihui/rmarkdown/books.html) package. --- <img src="Consolidating_Analysis_and_Dissemination_of_Research_Using_GitHub_files/figure-html/pages_to_book-1.png" width="1508" style="display: block; margin: auto;" /> --- <br> <img src="Consolidating_Analysis_and_Dissemination_of_Research_Using_GitHub_files/figure-html/example_workbook-1.png" width="2456" style="display: block; margin: auto;" /> --- # Implementation + Book can be hosted via GitHub Pages by following the approach outlined in section 6.3 of Authoring Books and Technical Documents with R Markdown [GitHub Pages](https://bookdown.org/yihui/bookdown/github.html) [ 6.3 of Authoring Books and Technical Documents with R Markdown](https://pages.github.com/) --- # Considerations This approach may be particularly useful: + For analyses that are **iteratively updated** and **shared** (e.g., a workbook for state data analysis; annual psychometric analyses) -- + As a space to share final analyses, after all the messy, iterative work has been done --- layout: true class: inverse, bottom --- # Supplemental Materials --- layout: false # Some Orienting References + Educational Data Mining ([Romero & Ventura, 2010](https://ieeexplore.ieee.org/document/5524021); [Baker & Yacef, 2009; Pena-Ayala, 2014](https://jedm.educationaldatamining.org/index.php/JEDM/article/view/8); [Dutt, Ismail, & Herawan , 2017](https://ieeexplore.ieee.org/abstract/document/7820050)) + Learning Analytics ([Ferguson, 2012](http://oro.open.ac.uk/36374/1/IJTEL40501_Ferguson%20Jan%202013.pdf); [Verbert et al., 2013](https://journals.sagepub.com/doi/abs/10.1177/0002764213479363)) + Data-Driven Documents ([Bostock, Ogievetsky , & Heer , 2011](https://ieeexplore.ieee.org/abstract/document/6064996)) + Data-Driven Decision Making ([Marsh, Pane, & Hamilton, 2006](https://www.rand.org/pubs/occasional_papers/OP170.html)) + Data Visualization + Classics ([Cleveland , 1994](https://www.amazon.com/Elements-Graphing-Data-William-Cleveland/dp/0963488414); [Tufte , 2001](https://www.edwardtufte.com/tufte/books_vdqi)) + Modern ([Few, 2013](http://www.stephen-few.com/idd.php); [Nussbaumer-Knaflic , 2015](http://www.storytellingwithdata.com/book)) + State Report Cards ([Curl & Peltzman , 2017](http://www.ccsso.org/sites/default/files/2017-11/CCSSO%20Reporting%20Best%20Practice%20Resource.pdf))