Community Event Guide: Evening Chatbot-Sprint

Hey everyone,

If you’re new to organizing community events, check out our Community Meetup Guidelines and Event Kit.

If you’re a meetup organizer and you’d like to host a chatbot-sprint, we really appreciate your help to spread the message about Rasa, and we want to help you get started by sharing useful tips!

Guide for an evening event

If you only have a single evening then you should recognize that time really flies. You’re only going to have 2-3 hours at your disposal so it’s best to keep the scope small. You probably won’t be able to explore all of Rasa’s features, and you will need to be mindful of time constraints while getting everything installed and set up on your computer. (which is especially true if you may have some folks who are new to Rasa.)

Here are essential resources to be aware of before your event:

  • We’ve made videos that will guide you through the installation process from scratch: windows/ubuntu/mac. Make sure to share them with your fellow community members beforehand!
  • A lot of events and workshops are available to watch on our YouTube live-coding a chatbot from scratch. You could share an example before your event with your attendees, this will allow them to familiarize with the format and prepare with questions. You can find one such example here.

If you’re planning to do a single event where you start the workshop project from scratch, then it may be best to focus onrasa init and start adding a few intents to your bot. Here’s a list of things that you may want to include:

  • A typical example being used it showing a photo of a tiger, but you can even start as simple as training your bot to tell a joke, and show how the bot can continue to tell jokes until the user mentions that they are happy.
  • Adapt the assistant so that it can handle frequently asked questions. Source a FAQ page from the internet, and see if you can make the assistant answer the questions appropriately.
  • Adapt the assistant so that it is able to reply to chitchat. Things like “what is your name?” and “what is your favorite food?” might be fun to add.
  • You can also translate the rasa init application such that it also works in another language.

Rasa Beginner Friendly Repositories

If the group is already somewhat familiar with Rasa then you may also consider spending an evening collaborating with an open source chatbot or contributing to a dataset. We’ve open sourced a pokemon themed chatbot and there’s also a nlu-dataset repo that you’re more than welcome to contribute to. If you’re planning on adding a feature, let us know on the GitHub issues beforehand. That way we can prevent double work for the same feature.

Guide for multiple evening events

If you’re planning multiple events on a regular basis (say, once every month) then you’re able to build something over a longer period of time which means that you can think of something more elaborate. You could contribute to an existing open source bot, but you can also start working on your own!

  • You could make a chatbot to support your own meetup group. One that is able to answer questions about the group such as past events, future events and about what the meetup is about. You can create a basic bot during the first event and you can host it during a second one.
  • You can also organize multiple events to explore increasingly advanced Rasa features. The first event may be just simple intents but subsequent events may involve custom actions, entities, forms or exploring the machine learning pipelines.
  • If you’ve got a group of experienced developers it may also be fun to start with Rasa-X. Get it deployed first and add features from there. The benefit of this approach is that you can share your assistant more easily and immediately get a feel for what it is like to run it “in production”. You can also make it available in a group slack. You can find the installation guide for Rasa-X here.

New to organizing community events? Check out our Community Meetup Guidelines and Event Kit to help you get started!

1 Like