six-hackathon

Authors: Betty, Seb, Nicolas, Alex @inpher.io

Bid Anon- A competitive bidding platform with privacy and insight

Description: Bid Anon is a privacy preserving open bidding platform that matches borrowers and lenders with rich statistical data and post transactional verification. Existing lending marketplaces are based on flawed models with limited data sets that either require you to publish personal information, use a ‘trusted’ third party broker, or rely on unsubstantiated user reviews. Built on advanced cryptographic technology, Bid Anon provides a zero-knowledge platform wherein anonymized data derived from financial information, social profiles, geolocation services and other public and private sources is used to rank borrower profiles through machine learning algorithms and provide a marketplace for lenders to bid on loan opportunities. Private information is only shared directly between the parties after the borrower accepts a bid in order to validate the transaction. Loan contracts can be traded through our backend blockchain integration for decentralized secondary market opportunities.

Value for the borrowers (sellers)

• Better deals through a large lender marketplace

Value for the lenders (buyers)

• Access to more opportunities through an online borrower marketplace
• Mitigated risk with insight into comprehensive borrower data

Opportunity for secondary markets (traders)

• Decentralized trading platform enables loans to be packaged and resold

Prerequisites

• Node.js
• MongoDB
• Bower

Getting started

Prepare environment

# Launch Mongodb in dedicated terminal
mkdir mondata
mongod --dbpath mondata

# Clone Repo
git clone https://github.com/Inpher/six-hackathon.git

# Download dependencies
cd six-hackathon/bidme
npm install && bower install

# Start node server
node server.js

Start web application

=========>>>> currently only Chrome is supported! <<<<==========

Open Browser: http://localhost:3000

Register as a bank

Under the hood

Basic Idea

The idea is to match borrowers with lenders anonymously based on objective criteria. Once the match has been done, the transaction is settled and identities are exchanged. This means that all sensitive information is processed on the client and only anonymized data is stored on the server. The private data is dumped in an encrypted blob and only made available to the lender once the match is completed.

Creating and Signing Contracts

Once the lender accepts a loan proposal (bid), the identities are revealed and a contract is created. The contract is cryptographically signed by the borrower, the private data (including identities) is encrypted with public keys for both the borrower and the lender (on the borrower side) so that it remains invisible. The contract is then sent to the lender (the bank) to reveal the private data (after decryption) and sign the final contract with the bank's signing key. The server is capable of verifying the integrity and the authenticity of the contract without having the private information (a blind signature).

Testing signing contracts in the javascript developerconsole

You can do unit tests for signing contracts from the console via the following steps:

• Login as a user (borrower) from the browser
• From the Chrome browser javascript console, run the following function:

> testEncryptedContractsGeneration()

You will see three outputs:

• the initial contract generated, signed and encrypted by the borrower as it will be stored in the server database
• the version of the contract received and decrypted by the lender (for review)
• the final version containing the signature of both the borrower and the lender.

You can see the test code in

$PROJECT_HOME/bidme/public/scripts/src/contractscheme.js

Financial Score

Sample financial information is located in:

$PROJECT_HOME/client-data/findata/

The data is pulled from the OpenBank project. The score is currently computed based on simple heuristics but with a large sample size of users we could create more detailed statistics regarding the financial capabilities of the user and match it with other census data such as average income for a country/region.

The following script shows a PoC how this data could be accessed from the app itself and not by uploading a file through the OpenBank rest api.

$PROJECT_HOME/retrieveTransactions.py

Social Score

Using the facebook graph API. The score is again calculated based on some heuristics of the following parameters:

• Age
• Education
• Work

Future work and improvements

Decentralized Trading Platform for Loan Contracts

Once the contract has been created, it can be traded between borrowers or lenders (as bonds are traded) in a decentralized way using blockchains. Suppose that a borrower has a mortgage contract (with a loan amount or "par value") and a maturity date. If that borrower wants to sell the loan contract to another borrower without going through a centralized brokerage scheme, one can represent the loan contract via Ethereum smart contract and securely change the ownership of the contract (the borrower). The contract buyer could then pay with cryptographic currency (bitcoin), again in a decentralized manner.