This repository is version of https://github.com/HigherEdData/Du-Bois-STEM that is designed to running entirely within the user's web browser with out needing and external server. It is powered by the JupyterLite project.
There is a GitHub automation included in this repository that build and deploys this repo as a GitHub pages site at https://highereddata.github.io/dubois-jupyterlite/
When developing this content, be sure to refresh your browser, clear your cache, and use the Help->Clear Browser Data to ensure you are looking at the most recent version of your content.
If you need to add an additional Python or R library to the software evironment, add the library to the environment.yml file.
Place any data files, notebooks, or images that you want to make available to learners in the content/
directory of these repo.
This README.md file is also automatically copied into the content directory when the site is generated.
Browsing to https://highereddata.github.io/dubois-jupyterlite/ starts a JupyterLab like interface showing a file browser and multiple notebook noteboks.
When sharing links to students, providing a direct URL may be preferred. To find this direct link, use this alternative URL:
https://highereddata.github.io/dubois-jupyterlite/tree
and then open the notebook you want to share. Copy the URL in the address bar.
Examples:
- https://highereddata.github.io/dubois-jupyterlite/notebooks/?path=notebooks%2Fpython_popmap_dubois-updated.ipynb
- https://highereddata.github.io/dubois-jupyterlite/notebooks/?path=notebooks%2Fpython_time_series_xy_dubois.ipynb
The argument of ?path= is relative to the contents/ directory of this repository.
When loading data (such as with a command like pd.read_csv(url) you may encounter an error involving the
the acronym CORS (Cross-Origin Resource Sharing). This is security feature of modern web-browsers.
One solution to this issue is to ensure any files (e.g. .csv or .png files) that you need to reference in these notebooks
are also included in the this repository. Then instead of using a URL, you can reference the relative path.
If the notebooks are in the directory contents/notebooks/, then the relative path to various files could be
Examples:
../data/data.csv../readings-images/original-plate-02.jpg../data/us_state_2020/us_state_2020_reduced.shp
Another solution is use URLs that have CORS configured to accept request from all origins (e.g. https://raw.githubusercontent.com)
Reference: jupyterlite/jupyterlite#729
JupyterLite attempts most of the functionality of JupyterLab instance that normally would require a server but running only in the local web browser.
While many libraries and modules have been successfully made to run with JupyterLite, there are some functionality which not (yet) work fully with JupyterLite. The Python and R libraries used in these Du Bois notebooks work as intended but there may be compatability issues if additional libraries are needed. The JupyterLite ecosystem is under active development.
This repository publishes data and Jupyter Notebooks for the STEM Data Visualization and Du Boisian Methods learning modules. The lessons (we use the terms module and lesson interchangeably) are designed for learning how to create scientific data visualizations in in multiple programming languages. The modules use examples from charts created by Black social scientist W.E.B. Du Bois and a diverse team of collaborators for the 1900 Paris World Expo.
The Du Bois team's visualizations were state of the art in 1900, using most of the major chart types still employed across the sciences and engineering today. The Du Bois charts provided some of the first widely accessible scientific analyses to refute false, biologically based theories of racial inequality. For their innovative analyses, beauty, and historical importance, the original hand drawn charts are preserved in the U.S. Library of Congress.
The modules are published in separate "lesson language" websites for your programming language of choice. The websites were produced using the Data Carpentries workbench for open source Github lesson sites. You can navigate to our lesson sites here:
- STEM Data Visualization and Du Boisian Methods with R
- STEM Data Visualization and Du Boisian Methods with Python
- STEM Data Visualization and Du Boisian Methods with Stata
- STEM Data Visualization and Du Boisian Methods with Pen and Paper
Each lesson language site is divided into separate episodes. Learners and instructors can select the the episodes that best suite their prior programming experience and learning objectives. All but the final episodes are designed for learners with little re no programming experience.
The first episode of every lesson site starts with the social and scientific context in which Du Bois created the visualizations. The lesson then introduces key chart types for different kinds of data and visualization best practices. Multiple episodes are then offered for using the lesson's language (R, Python, Stata) for recreating or adapting selected Du Bois charts with modern data and STEM field related data. Chart recreation eplisodes include Du Bois bar graphs and statistical maps.
The lesson sites are designed so that instructors can teach the modules from the site. Learners can also use the sites to learn their content independently and asynchronously. For instructors, we find that the sites are pedagogically most effective when used as lecture notes and learning activity instructions. We do not recommend lecturing with a screen share of the lesson site or projection of the lesson site. This combination of text, activity prompts, and verbal narration tends to exceed effective cognitive loads for learners.
Where appropriate, we instead provide images (like charts, diagrams, and photographs) that can be opened from the lesson site in separate browser tab for display. We also provide links to google slide files with all images for each episode. You can copy edit the google slide files suit your particular teaching practices.
Episode specific learning objectives and questions are noted at the beginning of each episode. Overall learning objectives for entire modules are:
- Understand the social and scientific context of visualizations as a process of scientific discovery (observation, hypothesis formation, data collection, analysis).
- Practice creative and visual thinking as valuable methods for scientific discovery and communication.
- Comprehend major chart types (pie bar, bar chart, line chart, statistical map) and their suitability for different levels of measurement and multivariate relationships.
- Apply visualization best including accessible design and data story telling.
- Create and modify a Du Bois chart using R to experience the value of coding to reproduce and adapt data visualizations in STEM.
Our coding interactives with Jupyter Notebooks can be accessed from the Du Bois Cloud using any web browser with no installation required.Links to the associated coding interactives are provided on each lesson site. But you can also navigate directly to them here:
- Literacy Bar Graph with R Notebook
- Literacy to Biodiversity with R Notebook
- Literacy Bar Graph with Python Notebook
- Black Population Map Python Notebook
The STEM Data Visualization and Du Boisian Methods learning modules were created by a collaboration between STEM researchers and instructors at University of California Merced, Princeton, and Fisk University where Du Bois earned his first undergraduate degree. Development and testing of the modules is funded by the National Science Foundation.