Design Proposal

Seth Corker
Benevolent Bytes
Published in
2 min readDec 18, 2016

--

Twitter and other social networks are places for users to share information with their friends and other intended audiences. These sites are intended to connect users and focuses on people as individuals. My idea to is to use publicly available information from social media sites in conjunction with other sources such as news headlines to compose content for a social media bot. Divided into each country/region, a bot would summarise the general feeling of the that region in a machine generated post. The post’s content would be based on the popularity of posts in that region and would be given context through matching current events. This would personify a country/region so someone could understand in a concise manner, how the country ‘felt’ and what a major talking point is.

Extension — AI Bot

To extend this idea further, I would like to explore creating a more interactive experience in the form of a chatbot. This would reside on a messaging platform such as Facebook Messenger or Telegram and would interact directly with users. A bot would be created to represent a country/region and would be capable of interacting via messaging for the purpose of gauging how the bot felt or responded to questions and thus gain insight into the country/region the bot is representing.

Deliverables

Stage 1 — One Way Interaction

A bot will compose messages based on news headlines and tweets from the region and will update its status incrementally.

- Social Network API Access — Intelligent post composition

These three requirements will allow for the first stage to be complete and offer information to be used in Stage 2.

Possible core technology used to realise this idea could include:

Stage 2 — Two Way Interaction

The bot will be interacted with directly by a user and will respond to some set of questions based on the news information it knows and the general ‘feeling’ of users in the region the bot is representing.

- Interactive Bot on Messaging Platform — Incremental Learning Based on User Input — Persona Influenced by WiderSocial Network

This stage will be more challenging but will the most interesting wot work with.

Possible core technology used to realise this stage could include the following:

  • ChatterBot (Machine Learning conversational dialog engine)
  • wit.ai (Natural Language Engine)

Stage 1 and Stage 2 are not co-dependent and one stage could be completed independently to accelerate the project or focus on particular functionality. I am very interested in completing Stage 2 because it offers the most interesting challenges.

Title Image: Source

--

--

Seth Corker
Benevolent Bytes

Sharing ideas, facts, and opinions about programming. 📚 Learn more React and JavaScript. 🗺 Learn web development tips and tricks 🛸 Discover something new