Cross-language analysis of U.S.-Russian relations via Twitter (Task 3) or early Alzheimer syndrome detection in speech analysis (Task 9)

Cross-language analysis of U.S.-Russian relations (Task 3)

Tweets from US democrats (or just anyone actively opposing Trump's presidency) and Russian conservators (or anyone supporting government) are collected and brought to a common denominator via translation-based and non-translation cross-lingual methods. Resulting corpus would be fascinating to research through several perspectives: for example get the political views through speech analysis, comparison of similarities in the vocabulary between people of similar and different political views, etc. Information can be retrieved through the Twitter API (Search or Stream, using Tweepy library).

Early Alzheimer syndrome detection in speech analysis (Task 9)

Research showed that it is possible to detect early signs of Alzheimer Syndrome in speech. I propose to train a model, which will learn to do it (and later make an app, assisting people with the high risk of Alzheimer). Data can be extracted from the Talk Bank (DementiaBank) or, for example, Dementia Diaries website. The app could prompt their users to talk about their day and using speech recognition module could analyze the possibility of Alzheimer and changes overtime.