Story Trident

• 

  Homepage

• 

  A story

• 

  A Branch

• 

  Machine Learning Diagram

Inspiration

Have you ever been reading a story and not liked the ending. Or writing your own story and have been torn as to how the plot should develop next. Then our application provides a solution for you.

What it does

Story Trident allows you to share your stories online in a community building environment, however as someone or even yourself is reading you could decide to fork the story and, for example, write a new ending, incorporate a new side quest, or supplement the character development; thereby creating two possible paths for the plot line to follow. Then next time someone is reading, for example, 'Tales of King Arthur' eventually they will hit a point where they can continue with 'Tales of King Arthur' or for example read your addition 'Lancelot's side journey'. This allows for an online community to proactively collaborate, disagree, and experiment with maybe interesting, powerful, or controversial topics and ideas to build a more fascinating, more elegant and more entertaining story. We keep a running count of the number of forks a story has which may indicate the story was originally very poorly written and everybody wants to make it better or that the premise it used was so unique and intellectually stimulating that everybody wants to write their own version. We also allow users to up-vote and down-vote branches of stories to indicate feedback for the writer and future readers. This voting is also how we determine the best, most entertaining stories.

How we built it

We used node.js, mongodb, and Google Cloud Compute for the back end services and react.js for the front end. We used Microsoft Azure for machine learning and trained on Project Gutenberg data.

Challenges I ran into

We used Microsoft Azure Machine Learning Studio to analyze what a particular user likes and the rest of our stories to suggest different stories for them. We had an extremely difficult time in getting a machine learning algorithm to be able to categorize short stories into genres and struggled to achieve a significant subset of literature data to train on.

Accomplishments that I'm proud of

We got our trained machine learning algorithm to work and identify genre from text with 89% accuracy. We got the back end up and running allowing us to write and fork our own stories.

What I learned

We learned a lot about machine learning, Microsoft Azure and even Google Cloud Machine Learning during the process. We learned react.js, mongodb, and Google Cloud Compute.

What's next for Trident Stories

We hope to leave a server live for some time to finalize the basic features, use it for ourselves and invite friends and family to check out our handiwork. If it is well received we may continue work on it in order to build a friendly platform for a hopefully equally amicable internet community looking for this kind of solution and or environment.

Built With

• azure
• gcp
• mongodb
• node.js
• react.js