Project Glass

• 

  Overview of the Architecture of Project Glass

Inspiration

The system was Mostafa's idea. In a world with transparency has become more of a public concern in recent years, it was important to provide a medium to hold charities accountable for the money they received in good faith, and to encourage people with the means to donate.

What it does

Project Glass is a system that uses BlockChain technology to track donations given to charitable organizations that have opted-in. Each donation is given a unique "tracking key" like the kind you get on parcels to track the status of deliveries. Donors can then lookup their donation on the Project Glass website to see exactly where each dollar ended up.

It also provides suggestions for where it is best for the organization to spend money. This is driven by a machine learning algorithm that detects events on data collected on topics relevant to the NGOs in the network. The ML algorithm detects relevant events, which are then dispatched using PubSub+ to the Project Glass partner-organizations. The organizations would then be able to see a live feed of relevant data that they can use better leverage their short-term investments.

How we built it

We use Blockchain and a proprietary currency to keep track of every dollar spent. Each invested dollar is turned into a unit of currency and tied to a transaction id. The transactions of every dollar is then logged into the blockchain from the time it is deposited till the time it is sent to an external entity (such as another NGO, or if it was used for an expense). A person with a tracking id can use it to look up the final destination of every dollar that they have spent, which adds transparency as a result. Auditors can also use this information to verify the claims of the NGO expenditure by matching their bank transactions to what they claimed in the system, this makes their job easier.

We use data gathering, AI, and PubSub+ to generate and publish events. We have a data stream that we run a time-series based machine learning algorithm which detects events. The events are then sent over a PubSub+ topic which is received by the Project Glass service and used to drive suggestions for where it is best for an organization to send money.

Challenges we ran into

The main challenge was adoption, how do we make sure that this system can easily be adopted given the use of a new currency?

The solution is to use the new currency strictly for tracking investments dollar-to-dollar. This currency cannot be used or exchanged in any other context as it is only meant to be used to augment the existing financial system with traceability. In Project Glass, we limited the use of this currency to compiling transactions to the ledger and mapping individual investments with every contribution they eventually make.

Built With

• java
• pubsub+
• solace
• spring