Chinese Practice

The main product in this project is app.R, an RShiny app that displays semantic links between Chinese words in a given vocabulary list.

The interesting thing about Chinese is that its character structure uniquely lends it to root-word analysis, without the need for lemmatizing (at least, in the Chinese form). This will be obvious to native Chinese, but less so for Americans and heritage speakers like me.

Resources for network graphing

Katya Ognyanova: Network Visualization

Max Woolf: plotly in R

Francois Briatte: ggnetwork

Andy Chen: Coupled Events with Shiny

hlendway via RStudio Community: Observe click events

Carson Sievert via Gist: Using keys with click events

Carson Sievert: Interactive events with plotly

Data

Word and character frequency dataset: SUBTLEX-CH: Chinese Word and Character Frequencies Based on Film Subtitles

Prompt

1. I need to learn Chinese.
2. I have an existing list of vocab words to learn.
3. Duolingo is not advanced enough. I've completed almost every lesson.
4. Quizlet is not complex enough. I need to use the words in context, not just know what they mean.
5. I need unpredictable sentences. I can't translate the same few sentences. I would like some randomly generated prompts.
6. I need an app where I can organize the words in a 2D space and add tags/labels to make semantic connections
7. I need to be able to see common root words, whether computationally or by drawing connections

Initial Plan

*wishlist

Medium

• React app?
• RShiny?

Sentence practice (Duolingo style)

2/2023: All of these capabilities are much better accomplished using ChatGPT lol. Pausing development

• How to query my own .csv files? SQL?

• English sentence prompt

  • What kind of sentence templates to use?

    • write my own English templates?

    • get someone else's list of sentences to practice in other languages? where to find this?

    • draw verbatim from a children's book?

  • How to populate the template?

    • randomly mad-lib from the English column of my Chinese vocab list?
      • involves labeling my own list for parts of speech either by hand or using ML*
    • get someone else's list of common words and parts of speech?
    • try lemmatization and look at others' generators

• User input answer box and submit button

• Some kind of reaction, then next button

  • What kind of grading to implement?
    • no grading, just vibes
    • grading based on .csv lookup
    • grading based on Google translate* (would need API or something)
    • grading based on ML*

Flashcard sorting

• Login (user input) to store user-specific information in separate .csvs*

• read in csv (using what package?)

• maybe highlight common root words and connect them automatically*

  • tokenize (split on character), then count and sort by frequency
  • select characters above a percentage or absolute threshold of frequency (or both/either) - variable common is TRUE
  • (in this vein, could implement pre-sorting before manual organization allowed)
  • when click preview, you can show/hide connections between root words - should we label the meaning?
    • connection is curved line drawn automatically between positions associated with the cards where common is TRUE (what is the math behind this)
    • color/connections grouped by root word (separate column after identifying common roots)

• card class (could maybe bootstrap this)

  • properties: rectangle, word (front), pinyin + definition (back), x, y, boolean selected, tag list (or single value), sub-tag, boolean flipped, boolean common, root word list (or single value)

  • when click, deselect others and select this one

  • when touching selection rectangle, boolean selected is TRUE

  • when click and drag, move

  • auto-sort button to arrange by tag (including untagged)

    • need to figure out the math behind this - something like number of cells in the matrix determined by the number of tags total, the x and y position determined by number of cells and a counter in a loop

  • study mode

    • when click, flip
    • when ctrl + click, hide (?)

  • organize mode

    • when selected is TRUE and add tag button clicked or shift + T, add tag in tag column of data frame (user input box)

• save option to save data frame to csv (with username*)