Here's the whole story
Inspiration: Today's modern day dialers aren't smart enough so we decided to build one ourselves. Enter Dynamic Dialer, a dialer ahead of its times. Unlike traditional dialers which only take in your call duration to make predictions, it takes in more context. We feed it the time of the day, the day of the week (and plan to feed location as well) to it. For example on weekdays you're more likely to call your colleagues at work and on weekends you're more likely to call your friends who you party with. So feeding it more context helps it make smart predictions unlike traditional dialers which will always show as their top prediction, the person who has the highest call time in your logs irrespective of which day or where you are (office or near a party place :P)
We built it on Android and used an online Machine Learning Service as our backend. We feed in the person's call logs to the service and the service returns a list of people who you are likely to call based on the decreasing order of probabilities.
Challenges we ran into: Interfacing the Android application with the Online Service took us a great amount of time Accomplishments we're proud of: To see it working completely. Please come to our table and check it out :)
Whats next: Feeding it your location and syncing it with your events(Google Events/Calendar etc.) . For example I was more likely to call my InOut team mates this weekend or people who stay in Bangalore since I was travelling here for the hackathon
Log in or sign up for Devpost to join the conversation.