Input: Voice data Upload voice data to Watson cloud Watson Cloud analyzes data to determine:
- Tone
- Content
Web Input
- Take audio input
- Convert to correct format
- Run python script
Web Output
- Take input (GET request)
- Process input and add to mySQL database
- Return unique key that can access specific value
Web report
- Audio file playback
- All pertinant data
- PDF link (to download PDF)
Output report (PDF)
- Voice to text transcription of event
- What data contains
- Red flags
- Possible actions that can be taken (and why they should be taken)
- Maybe use LaTeX to make things easier??
Objectives:
- SpeechToText: convert audio files into text files
- ToneAnalyzer: analyze negativities from text files into numerical scores
- NaturalLanguageUnderstanding: analyze semantic features of text input
- Analyze text for "redflag" keywords
- May involve potential actions the user will take
- Generate report on the inputted audio/text file.
- Add progress bar for visualizations of each category.
Look for:
- Danger to Self
- Words or actions showing intent to commit suicide or bodily harm.
- Words or actions indicating gross disregard for personal safety.
- Words or actions indicating a specific plan to suicide.
- Means available to carry out suicide plan (i.e. pills, firearms present or available).
- Danger to Others
- Threats against particular individuals.
- Attempts to harm certain individuals.
- Means available to carry out threats or to repeat attempts (i.e.firearms, or other weapons).
- Expressed intention or attempts to engage in dangerous activity
- Grave disiblity
- Signs of malnourishment or dehydration.
- Inability to articulate a plan for obtaining food.
- No food available in the house or at hand if not in a house.
- Irrational beliefs about food that is available (i.e. it’s poisoned, inedible, etc.)
- Destruction or giving away of clothing to the point where the person cannot clothe themselves.
- Inability to formulate a reasonable plan to obtain shelter.