title | parent | has_children | nav_order | mathjax |
---|---|---|---|---|
Title of page |
Category |
false |
1 |
true |
If this is your first time writing a new LOST page, please consider giving the Contributing page a look.
INTRODUCTION SECTION
Remember that you can use inline math, e.g.
You can also render math in display mode:
The Introduction, Keep in Mind, and Also Consider sections should be LANGUAGE-AGNOSTIC, i.e. they should not be specific to a given language. Language-specific stuff goes in Implementations.
- LIST OF IMPORTANT THINGS TO REMEMBER ABOUT USING THE TECHNIQUE
- LIST OF OTHER TECHNIQUES THAT WILL COMMONLY BE USED ALONGSIDE THIS PAGE'S TECHNIQUE
- (E.G. LINEAR REGRESSION LINKS TO ROBUST STANDARD ERRORS),
- OR INSTEAD OF THIS TECHNIQUE
- (E.G. PIE CHART LINKS TO A BAR PLOT AS AN ALTERNATIVE)
- WITH EXPLANATION
- INCLUDE LINKS TO OTHER LOST PAGES WITH THE FORMAT [Description]({{ "/Category/page.html" | relative_url }}). Categories include Data_Manipulation, Geo-Spatial, Machine_Learning, Model_Estimation, Presentation, Summary_Statistics, Time_Series, and Other, with subcategories at some points. Check the URL of the page you want to link to on the published site.
Notes on implementations:
- Keep in mind this will NOT PRODUCE CODE OUTPUT FOR YOU, so if you want output to be visible you'll need to include it directly (and on that note, submit this as an .md file, not .rmd).
- Keep implementations fairly short - you're not writing a book chapter here! A single straightforward demonstration is often all you need. Extra demonstrations can be handy if there's an important alternate task you want to show. You don't need to show off every option.
- Ideally we will use the same data for all the examples across all the languages. So it's a good idea to convert your data into something universal (like CSV).
Commented code demonstrating the technique
LOST does have a code checker for some languages. If you have code split over multiple code chunks, and earlier code chunks are necessary to run later ones, give the chunk an ID like so:
x = 1
x = x + 1
Or if the code checker should skip it, say so!
read.csv('file-that-does-not-exist.csv')
There are two ways to perform this technique in language/software 2.
First, explanation of what is different about the first way:
Commented code demonstrating the technique
Second, explanation of what is different about the second way:
Commented code demonstrating the technique