DataDividend

Inspiration

Data leakage has become the new norm - and we stopped questioning it. Every app we install , every website we visit, every search we make is silent packed, sold and traded between companies we never agreed to work with . When the data is sensitive - health records, financial behaviour, location history - the consequences aren't just inconvenient. They are life- altering. Jobs lost, Loans denied. People targeted and through all of it , the person the data belongs to receive nothing .

What disturbed us most wasn't just the leakage but the power aasymmetry. Companies treat user data as an asset on their balance sheet. They decide who buys it , when it's sold and at what price. The user is removed from the transaction entirely .

We built 'DataDividend' because we believe that equation needs to be reserved. Your data should work for you . You should be deciding who sees it, what they see and what amount should they pay for it . And all this should be done without ever revealing your identity.

What it does

DataDividend is a privacy -preserving data marketplace built on the Midnight blockchain . It gives users the ability to :

1. Own their ddata profile - behavioural insights derived locally on their device, never upploaded rw to any server.
2. Set consent on their terms - choose which buyer can accesss which data segment .
3. Earn from their data anonymously - receive payments in DUST when buyers purchase verified proofs of their segment membership.
   1. Stay completely anonymous - buyers never see who the user is, only a cryptographic proof that the user belongs to a relevant segment.

The core insight is - buyers don't need your data. They need to know something true about you. An advertiser doesn't need your browsing history. They need to know you're a tech-interested, high-income user between 25 and 35. DataDividend proves that claim — without revealing the history behind it. The privacy isn't a feature. It's the architecture.

How we built it

The stack is built across three layers, each with a distinct responsibility: Layer 1 — Midnight Smart Contract (Compact) The core logic lives in a Compact contract deployed on the Midnight testnet. Midnight's unique private/public state model is what makes this entire system possible:

Private state holds the user's data profile, consent records, and earnings history — visible only to the user Public state holds proof issuance counts and buyer registry — visible on-chain

Layer 2 — API & Proof Engine (Yet to be added) A lightweight Express API bridges the Midnight contract and the frontend. The proof engine:

1. Evaluates the user's local data profile against hardcoded segment rules
2. Computes confidence scores per segment
3. Generates a structured proof certificate with public inputs only
4. Submits the proof to Midnight and triggers the payment transfer

Layer 3 — Frontend Dashboard (Next.js) Three screens, deliberately scoped:

User Dashboard — view your private segments, manage consent toggles, set price floors, track earnings Buyer Dashboard — browse available segments, see verified user counts, purchase proof bundles Proof Verification — verify any proof hash on-chain, confirm segment validity, confirm user anonymity.

Challenges we ran into

1. The Anonymity Paradox The first architectural challenge we hit was fundamental: if a user is anonymous, why would a buyer trust the data? In traditional data markets, identity enables verification. Our insight was that Midnight's ZK proof system inverts this — the proof is the verification. A buyer doesn't need to know who you are to trust that the proof was generated correctly. The cryptographic guarantee replaces the identity guarantee.

   1. ZK Circuit Scope A production ZK circuit for behavioural segmentation is weeks of engineering work. In given constraint timeline , we made a deliberate decision: build the correct proof structure with mocked circuit computation, rather than a broken real circuit. The proof output format, the public/private input separation, and the on-chain verification flow are all architecturally correct. The circuit computation itself is the one component marked for production implementation post-hackathon.

Accomplishments that we're proud of

1. A ZK Query Engine that thinks like a Data Buyer We built something that didn't exist before this hackathon : a zero - knowledge query interface where a data buyer can write SQL - like queries against encrypted user datasets - without ever accessing the underlying data.

2. A real - time cost enstimation query complexity Before a buyer executes any query , DataDividend computes: -> Base Fee - fixed cost per query -> Complexity rating - how computationally expensive the ZK circuit is -> Estimated reach - how many nodes the query will touch -> Total cost in DUST

What we learned

While working on this project, we learned :

1. The business model of data marketplaces has always been broken at the identity layer. ZK proofs don't just add privacy — they fix the fundamental value exchange.

2. Scope decisions made in the first hour determine whether you ship. Every feature we cut before building saved us two hours we would have lost debugging it.

3. Midnight's private state model is architecturally unique — it isn't just privacy, it's selective provability. That distinction changes what products are possible.

What's next for DataDividend

-> Production ZK circuit implementation using Midnight's native proof system -> Browser extension for authentic local data collection -> Sybil resistance mechanism — proof-of-humanity integration to prevent fake profile generation -> Dynamic segment marketplace where buyers can define and price custom segments -> Mobile-first consent dashboard

Built With

• compact
• css
• javascript
• typescript