Data Diary With Chatbots

09.2016 _ 11.2016 / boston

How well can a chatbot converse with a human user?


By applying sentiment tagging to the conversations with four commercial chatbots (Siri, Alexa, Google Now, Microsoft xiaoIce ), this animated data visualization constructed an “empathy map” that grounded the idea for my M.F.A thesis which regards virtual personal assistant as an affective interface.

The chatbots engaged different techniques to provided empathetic responses to tones and emotions they detected from the user through the turn-taking conversations, though sometimes still failed as babblings of nonsense.












The bi-product of this fun experiment is my MFA Thesis book which talks a lot about Embodied Interaction and how to design for Conversational User Interface. Not super mature writing but you could still take a look at MY MFA THESIS since there are plenty of concept maps I created for CUI and some discussion about how to deal with bot as human.

_ development tool:
AlchemyLanguage, node.js, jquery, d3.js