Journalyze

playing around with Speech to Text Technology

A speech to text transcription tool which allows you to explore your own words in a new way.

How it started

The technological field of natural speech analysis has made great strides in recent years. We investigated new fields of potential in this area. Speech to text Applications like Alexa, Google Home or Siri can make life more convenient. Even big companies pay for software to analyze customer needs, critical of that fact we asked us this question:

What if this important and sensitive data can be used to get a better insight into our own mind, and which profound interactions can we create to enable this process?

The Idea

We wanted to apply it to a self-reflection method, so-called "journaling". Here you recapitulate daily events to reflect on them. In this case Natural Language Processing (NLP) helps you with that.

↳ User-testing: Luca records a journal for our research
↳ Speech analysis of Lucas data set with the IBM-Watson NLP API

Design Process

The application should enable the user to comfortably explore. So we tested several navigation patterns to come out with solutions that suit individual preferences. There is a quick and fast way, but also a deeper layer of information the user can dive into. Because we work with sensitive data the design has to have a trustworthy appeal. The Wording, visual-design and motion design adds this profound character to the product.

Transcribing

Simply record journal entries, which are then transcribed and analyzed for you. Smart formatting converts your text legible.

↳ Talk about your day!
↳ A quick summary of the entry you're just did

Search & Discover

Discover clues or search through your entries for key moments, emotions and everything else that is important to you.

↳ Slide trough highlighted snippets of your journal
↳ Dive deeper into your transcript and search for sentiment

Manifestation

Find important insights or motivating quotations in your journal entries and save them in the notes.

↳ Mark important text an process it
↳ Safed text passages will be saved and pushed into a feed

Credits

Team: Vincent Fischer, Nico Brand
Consulting: Prof. Hans Krämer, Ann-Katrin Spörl
University: HfG Schwäbisch Gmünd, Winter Semester 2017

back on top next project previous project