Seamlessly tying together data exploration and analysis, technical blogging, and academic paper writing.
The vision:
- Write about your ideas, hypotheses, and experimental setup with syntax that stays out of your way with Markdown
- Put executable code in code blocks to load your data and analyze it
- Display generated plots and tables inline
- Output to PDF in the correct format of your target academic publication to see how the final paper will look as you go
- Output to HTML in a technical blog style for easier sharing
- Automatically make the tables and plots interactive so collaborators and strangers can manipulate and play with your data in new ways.
- Pandoc: generate HTML or Latex from Markdown (or a slew of other markup languages)
- extensible so we can build filters specifically for common use cases
- R's
knitr
package: actually execute R code blocks and display generated plots inline! - Plotly: take ggplot plots and automatically make them interactive, letting you drill into the data
- Currently Latex/PDF generation is working pretty well.
- Still trying to work out kinks in Pandoc/Knitr's mechanisms for references and labels
- Would like to add
\autoref
functionality, seems like Pandoc has some efforts in the works but nothing ready
- HTML generation currently poor because we're generating PDFs for plots and not converting them
- use Plotly when generating HTML to fix this problem