New speak

• 

  returned results

• 

  one of the returned article

Inspiration To engage people more into news fetching by interaction.

What it does The API takes keywords from user via speech. Thereby it searches from the database the most relevant ones by creating a ranking of the articles based on the given keywords as input.

How I built it At first 50k articles are fetched from Twipe API. The unique word occurrences excepting the articles, prepositions were counted and formed into a dictionary here and stored as a database. The Google API was used to fetch user input in form of speech. The user input is fed to the backend through AJAX and then processed by the ranking algorithm. This ranking algorithm ranks the most relevant articles from the database. This ranking algorithm calculated the frequency count of the user input and inverse frequency of the terms. The articles with images were given priority.

Challenges I ran into Effective collaboration.

Accomplishments that I'm proud of A working prototype

What I learned Tango with Django!

What's next for New speak IPO ;)

Built With

• ajax
• django
• google-speech-recognition-api
• javascript
• jquery
• machine-learning
• matlab
• microphone
• python
• twipe