Context

Work in Progress.....

Reduction  is based on feedback I got from a different product I was working on. Many ESL students talked about how overwhelming it was to read through an article in English, and how their current responsibilities made them feel really busy. Many of them wanted a solution to simply help them understand the ideas faster -- help save time while reading.

The goal of  Reduction  is to try to be like a multilingual Blinkist or Sparknotes on any text the user wants to read.

Blinkist and Sparknotes are both amazing products. One potential problem with Blinkist or Sparknotes is that, if the user only reads through them, the user does not know what information they are missing. The user generally trusts the prestige of the website and company that the text is successfully reduced. It is harder to trust a computer.

To try to address this need within a computer-based product, I created two cool metrics: an Information Retention score and a potential missing concepts list. The Information Retention score tries to find how much core information is retained in the article.
If there are very important concepts from the main text that are not found or discussed enough proportionally in the new version, they are highlighted for you, the user, to explore further.

The other core idea is creating customizable lengths for reduced texts. In an early prototype in Python, I had versions between 5% and 95% word count of the original text. Currently, I only have one or two basic lengths included, but I will be working to figure out how to transfer the algorithm appropriately into Javascript.

I hope you enjoy the prototype! Message me @BrianSwanberg on Instagram or email me at brianswanberg@gmail.com if you have any questions.