Speak Journal

Inspiration

I journal every morning in a handwritten journal. This is a very popular habit. However, it takes a very long time to search/analyze a handwritten journal. Electronic journals are great, but in order to write you have to turn on your phone/computer and you will be bombarded by distractions and notifications. Speak journal gives the best of both worlds.

What it does

All you have to do is Speak! Speak Journal will ask a few simple questions and listen to your responses (through the microphone). No screen is required. All answers will be logged and automatically tagged with keywords. The tags are found by looking through the responses for common keywords and sorting them into categories (happy, sad, stressed, excited, hot, cold, hungry etc). Speak Journal will look for patterns in recent responses and communicate these to the users. The journal is now electronic and can be analyzed/searched.

How we built it

Everything is written in python using SpeechRecognition and pyttsx3 packages for speech recognition and text-to-speech respectively. For simplicity the journal is stored in a text file, as well as the tags to the responses.

Challenges we ran into

I wanted to run the Speak Journal on the Raspberry Pi. Unfortunately I ran into many problems working with Microphone input to the raspberry pi. At first the raspberry pi couldn't recognize the microphone, and when I solved that the microphone was picking up lots of background noise.

Accomplishments that we're proud of

I built something that I will use every day! It might not be flashy, but it is useful to me!

What we learned

I learned about the python SpeechRecognition and text-to-speech packages. I also learned that the raspberry pi can be tricky to work with.

What's next for Speak Journal

The first step is getting Speak Journal on a portable device (raspberry pi or iPhone). Then, I want to add more functionality and "brains" to the Journal. In particular, the journal should be able to ask more probing questions to encourage insightful responses. Also, I want the journal to be smarter in how it analyzes responses and have the ability to look for long term patterns. For example, Speak Journal should ask how much sleep the user got and look for patterns in sleep length and mood.