Quest Data Engine

Provides composable data layers for developers to bootstrap their dApps while enabling users to store, manage and monetize their decentralized data with on-chain smart contracts.

To build a social application, a developer should only focus on user experience, user acquisition, and content moderation rather than worrying about building and maintaining the infrastructure. Quest supports developers to build new applications through APIs and SDKs. Although the experience is seamless to developers, we want to highlight our architectural design that makes Quest composable, and highly scalable.

Sufficiently decentralized data infrastructures that help users manage their portable and self-sovereign social data at scale. The data are represented in verifiable credentials and developers can safely rely on the infrastructures to write and update millions of records.

Data Modules​

Users generate various data when they are engaging in on-chain and off-chain platforms. We categorize them into the following three components.

Actions Module

Actions are activity-related data performed by the user. These include -

• Proof of Contribution via attendance at events, donations, completion of tasks

• Proof of Participation on Twitter spaces, raffles, competitions, whitelists, events

• Proof of Engagement via likes, comments, posts, reshares, messages, reactions

• Proof of Activity via clicks, views, view time, visits, scroll length, and more

• Proof of Transactions via wallet send and buy activity.

Affiliations Module

Affiliations are entities that the user is associated with.

Quest Graph Engine

Quest Graph Engine indexes, aggregates onchain and off-chain data sources. Developers could easily leverage the indexed data and the recommendation models for user acquisition and user retention.

Quest Graph Engine indexes, aggregates onchain and offchain data sources of a user to create a holistic view. Developers could easily leverage the indexed data and the recommendation models for user acquisition and user retention.

Data Indexing

There are three data sources in our indexing system: blockchains (e.g. Ethereum, Solana, BNB Chain), decentralized stores (e.g. IPFS, Arweave), and off-chain platforms (e.g. Foundation, Rarible, Twitter). Its general philosophy is to index data that could be potentially reused in various social contexts. For example, connection data on existing platforms like Foundation, POAP token for each address.

Data indexing is composed of three components: data processing, data aggregation and data querier.

Data Processor​

The processor connects to different blockchain full nodes (e.g. ETH, Solana) to watch real-time contract event data and run cron jobs to regularly fetch data from off-chain platforms (e.g. Twitter, Context, Foundation). It stores this indexed information as structured data in relational databases such as MySQL and PostgreSQL. These data are indexed for long-term persistent storage to avoid re-processing penalties.

Data Aggregator​

The aggregator aggregates structured data from relational databases into graph representation and stores that aggregated data in graph databases. Storing data in a graph database gives developers more query capabilities and makes graph-based data analysis and recommendation possible.

Data Querier​

The Querier serves three main purposes: providing identity data, providing relationship data, and connection recommendations. It retrieves data from graph databases, gets recommendation results from the Recommender, and returns data in the way requested

Recommendation

The recommendation engines are meant to lower the barrier of launching applications even further and increase access to high-quality ML models for new developers. We need to equip developers with the tools needed to build decentralized social applications that are truly competitive with their Web2 counterparts.

Types of Models -

1. User Recommendations

   1. Recommended Accounts: Get a personalized list of EOA (Externally Owned Accounts) to target based on a user’s interests and what similar users likes.

   2. Recommended Contracts: Get a list of contracts to target based on a user’s interests (ex. NFT ownership) and/or who are followed by similar users who have similar interests as the input user.

2. Item Recommendations

   1. Reward Recommendation: Get a personalized set of recommended rewards based on a user’s interests and what similar users have engaged with. This can look like recommended specific NFTs, tokens and/or games to engage with.

   2. Related Content: Get rewards similar to other rewards based on categories or tags along with user engagement. For example, getting recommendations for Bored Ape Kennel Club after opening the profile page of BAYC.

3. Feed Recommendations

   1. User Feeds: Feeds catered to a specific user / address' on-chain interactions

   2. Community Feeds: Feeds catered to a specific community's content & on-chain interactions. Get recent user activity related to a specific community (ex. Azuki holders, UNI voters, etc.)

Domain Feeds: Feeds catered to a specific domains' content & on-chain interactions (ex. Social, Defi, etc.) Get recent user activity related to a specific item (account, product, tag, post, video, etc.)