-
Happy text at main text entry interface.
-
Happy text video suggestion.
-
Sad text at main text entry interface.
-
Sad text video suggestion.
-
Alternative sad text video suggestion.
-
Angry text video suggestion.
-
Excited text video suggestion.
-
Fear text video suggestion.
-
Video suggestion for any other detected mood.
Inspiration
Mental health awareness is an important issue in our time. However, it's not always easy to change how you feel. We were inspired by the idea that an AI can detect emotions in a text to build this diary-like web app to help figure out one's own emotions and improve mood by recommending videos based on text.
We were also inspired by the now non-existent Youtube appication called moodwall, which recommends videos for different sentiments.
What it does
It analyses the text input and predicts your mood. Then it recommends a Youtube video to cheer you up if you're sad, to boost your mood if you're already happy and so on.
How we built it
We used HTML and css and javascript to build a web app. We used python to build server. Combining it with the IBM Watson's API to process sentiments in the given text, we then used thesaurus to find the antonyms of negative emotions to recommend youtube videos.
Challenges we ran into
Incorporating API, embedding videos into web app, how to recommend different types of videos, debugging
What we learned
Using API, Web and logo design, Python, HTML and javascript, Flask, Time management and team work ;)
What's next for MoodTube
This app can be combined with social media apps such as Twitter and Facebook to not deal with writing input texts and it would work by just scanning your posts and messages. It also can be used for personal development as well as by psychologists and therapists to analyze the mental status of their patients. The video recommending part can be improved by adding useful resources and personalized suggestions.
Log in or sign up for Devpost to join the conversation.