Data4.builders
Data4.builders provides massive user behavior data for developers without sacrificing users’ privacy. We are building a composable, universal user profile data layer that is high efficiency, privacy-preserving, and fits web3 paradigm.
Demo
About
**Data4.builders provides massive user behavior data for developers without sacrificing users’ privacy** **User behavioral data** is being stolen by tech giants, recklessly used, and even sold for their own benefit, media platforms are using this data to increase user retention, advertising platforms are using this data to drive deals, and in web2, it's a $2.8B market. Today, Data4.builders will take back the data from Tech giants and make it available to developers to create value without sacrificing users’ privacy. **Data4.builders is a composable, universal user profile data layer that has the following features:** - **High efficiency** - Data is a non-competitive resource that has more value if it is used more. However, web2’s data silo makes data only flow in separate tech giants. By aggregating all data into a layer makes everyone can use data to create value, which makes data more efficient. - **Privacy-preserving** - Users can control whether their data is used or not and their identity is kept completely confidential. - **Web3 paradigm** - Compared to web2’s DMP(data management platform), web3 has a new identifier other than cookies & device No., which is wallet address, and web3 data is a new goldmine for value creation. Data4.builders re-designed the DMP into the data layer and adopts it to web3. - **AI enabled** - All tags are generated using **LLM** from user's transactions, we can have massive user data just by typing in the address, without any kind of human curation.
How it's made
For the demo, we made a simple news platform that uses our data layer for **content recommendation**. **Technologies used:** - **Identity**: Next.id, Sismo - **Data query**: Airstack, Quick Node - **Wallet**: Wallet Connect - **LLM**: GPT-3.5 turbo - **Others**: Next.js, Vercel, MangoDB **Key components:** - **Identity aggregator** Identity aggregator gathers users' accounts in different ways and provides identities for Address to tag API to analyze. Using **[Next.id](https://ethglobal.com/events/paris2023/prizes#next-id-and-mask-network)**’s relation service, we can have a user’s Socials(Twitter, Github, Discord) from his/her address and ENS, etc. Using **[Sismo](https://ethglobal.com/events/paris2023/prizes#sismo)**, we can turn some data gem into tags and do not need users to connect several wallets. - **Address to tag API** Input an address, and output tags related to these addresses. Using **[Airstack](https://ethglobal.com/events/paris2023/prizes#airstack)**, we can query all the transaction history for a user and we use GPT-3.5 as the **LLM** to tag them, with **[QuickNode](https://ethglobal.com/events/paris2023/prizes#quicknode)**, we get all NFT collections related to the address. - **Content matcher & Content platform** This is just a demo and shows how data4.builders can be used in content recommendation scenarios. We use React, Next.js, and **[Wallet Connect](https://ethglobal.com/events/paris2023/prizes#walletconnect)** , also, we use **LLM** to get tags from the content and match them with users' tags.