charityX.ai

Merchant's view after selling
System architecture
blackboard discussion/designing

Inspiration

Ideally: When disaster strikes, people donate to charities and charities distributed donations to those in need

The reality is Most donations actually go to running the charities, in 2016 according to a forbes article, on average 85% of donations go to administration costs [1]. That means 3 years ago only 15 cents of every dollar donated actually reach the people who need it. This led to an effort to reduce the administration overhead, and charities have responded with the best and most trusted charities only spending up to 25%. However it’s still quite high with an average of 60 - 70%. Aside from these problems charities often have to work with local governments, and a good amount of donations are misappropriated or stolen. There is also the problem of unneeded donations, a handful of east African countries are actually banning clothing donations because it devastated their local clothing industry [2]

What it does

Our app is an innovative solution to all these problems! Our app has:

Negligible Overhead - By donating directly to people we shrink costs!
Automated Trust- Computer Vision with AutoML lets people easily verify their purchases. This feature also lets our app age backwards with new transactions making it better.
Need- Helps analyze what people really need by generating data for AI analysis, helping charities understand demographics
No internet required!- Thanks to the twilio API, users and merchants can trade with donated funds without ever needing to connect to the internet. Being able to trade goods cashless without needing an internet connection!

How we built it

We built two APIs: a core web API and an ML/Twilio API that share a single database. The core web api is built Django and is responsible for displaying disasters, donation requests, and accepting donations. The ML/Twilio API is responsible for Computer Vision verification and for the sms/mms interaction. The two APIs work together in updating the database. Stripe for payments

Challenges we ran into

Two very different APIs needing to interact with each other and a database led to a lot of requests issues we weren't familiar with

Accomplishments that we're proud of

1.Being able to trade cashless without needing an internet connection!

2.auto-AutoML - scraper/data preparer script that sets up automl for training just by providing classes and count for scraping.

3.smooth text message interactions with django

4.discovering cryptographic tricks for offline verification

What we learned

Flask, networking, twilio, automl (most of the tech we used), charity stats

What's next for charityX.ai

Make it more robust and secure. Talk with some business experts for possible improvements, and possibly a startup launch. There is great potential for this app to help a lot of people.