Thymio is an open-source educational robot designed by researchers from the EPFL, in collaboration with ECAL, and produced by Mobsya, a nonprofit association whose mission is to offer comprehensive, engaging STEAM journeys to learners of all ages.
This repository showcases a proof of concept I wrote for a webinar about the Thymio robot1 that connects to the Twitter API, fetches recent tweets where Thymio is mentioned, performs sentiment analysis on the content of the tweets, and activates the green Thymio leds if the result is positive, and the red leds if it's negative. Thus, it gives Thymio a kind of artificial emotionnal intelligence.
Further developments would include improving the quality of my code, obviously, but also bringing the concept of creating a bridge between live feeds via API's and actual robots reacting to this live information. Using Machine Learning on live feeds of information seems like an interesting path in giving connected objects a kind of autonomous and spontaneous intelligence.
The main program uses the following libraries :
- TDM client : Communication with Thymio II robot via the Thymio Device Manager
- Tweepy : Python library for accessing the Twitter API
- RE : Regular expressions operations
- Textblob : Python library for processing textual data
- A comprehensive guide to using Thymio from scratch can be found here : https://www.thymio.org/products/programming-with-thymio-suite/.
- In order to use the Twitter API, one has to create a developper account and request for API keys : https://developer.twitter.com/en/docs/twitter-api.
- A very useful guide to the process of creating a Twitter API developper account and setting up various twitter bots using Tweepy can be found here : https://realpython.com/twitter-bot-python-tweepy.
Here is an html version of the slides I prepared for the actual webinar about the Thymio robot : https://files.modulo-info.ch/markdown/presentation.html#slide1.
Note : this presentation was written in markdown via this tool : https://github.com/adamzap/landslide.
The source markdown file is visible here, for information : https://github.com/elliotvaucher/thymio_twitter/blob/main/docs/thymio.md.
The functions for sentiment analysis using textblob were taken from this tutorial : https://www.geeksforgeeks.org/twitter-sentiment-analysis-using-python/. The step by step guide for setting-up Twitter API keys comes from here : https://realpython.com/twitter-bot-python-tweepy.