Help.me - I always wished to build a convenient tool for complete proactive as well as reactive situations during disasters as during such critical situations, we need to access information quickly and all the information that we are provided with should be important.
NGScore - Looking at the uneven distribution of funding among the various non-profit organizations, I wished to come up with a solution that leverages machine learning to compare non-profit organizations irrespective of their magnitude in infrastructure because there are always small organizations which deserve more credit.
What it does
Help.me - Its a FB messenger bot that provides tips, weather and reactive measures such as directions to safest points, calling for helpline just at the touch of a button in a quick typing-free way.
NGScore - Uses Machine learning to provide a single metric score based on parameters of the organization irrespective of the size of it.
How I built it
I used Tensorflow to build the server that computes the score for NGScore and Chatfuel for Help.me to provide an intuitive UI.
Challenges I ran into
Hosting the server on an AWS EC2 Instance and making it publicly accessible.
Accomplishments that I'm proud of
After trying to build a huge system and failing due to technical glitches, I am proud that I was finally able to achieve something that creates value for real life users and can be used on the go.
What I learned
I learned project management better, How to evaluate an idea and build things that create value than just being fancy enough to make people go WOW.
What's next for HelpMe and NGScore
HelpMe can be integrated with NGScore as both of them can be made into a single platform web and android app for non profit organizations and corporate companies who wish to do CSR and also improve the algorithm for learning of NGScore to be complex for better evaluation.