Tips on creating your own corpus

[icebear]

TLDR: Evernote can serve as a convenient, free tool to build a personal corpus, improving automated selection of high-frequency words to study tailored to each student.

---

 

For awhile now I've felt challenged with how best to add new vocabulary to my flashcards/studies. While I've found tools like Imron's analyzer immensely useful, there always has been a little extra legwork to get the optimal utility out of them.

 

Problem

I read an awful lot of Chinese news articles and other shortish pieces each day. I also occasionally read a book or longer story. I listen to podcasts with transcripts daily. What new vocab do I commit to studying? Copying each one of these individually into a text analyzer to extract the most frequent words still leaves the problem of which to select from all stories in aggregate. If I even just choose one per source I quickly would be overwhelmed. Also, to me it makes sense to select from all articles at once, given a lot of the jargon will show up once per article but perhaps many times across them all.

 

Before figuring out my latest technique, I essentially opened Notepad.exe every morning and copied any text into it after I finished reading . Then I'd check that batch of text at the end of each day for the most frequent words within it. Unfortunately, this felt like a huge chore and also incomplete - what about overlap across days/weeks/years? I wasn't interested in building and maintaining a monster archive myself... and I have tried a few tools like Instapaper without finding them worthwhile (for this specific task).

 

Solution

After hearing about it for years I finally gave Evernote a try for a separate work related project. While using it with its Clearly web plugin I realized it could meet much of my Chinese archiving needs out of the box, for free. As I read articles I use Clearly to clean up often horrendously ad-heavy Chinese web sites, read, and when I'm done I will save to Evernote. I can easily add in ChinesePod transcripts, books, whatever. Once a week I export all my saved notes from Evernote into a consolidated HTML page, which I then copy into Imron's CTA. So far the process is far more convenient and the output seems much more useful.