Overcollateralized borrowers deposit funds into Sky and receive DAI stablecoin.

Sky is a non-custodial stablecoin protocol built on Ethereum. Sky (previously MakerDAO) was founded by Rune Christensen and launched in 2015.

SKY tokenholders own and govern the protocol. They currently take all of the total interest and PSM fees paid by borrowers (revenue). Sky’s investors include a16z, Polychain, IOSG Ventures, etc.

governance

github

twitter

discord

etherscan

website

Stablecoin issuers

Sky | Fees and revenues breakdown

# Sky Active users (daily) Methodology

## 1. Overview

Daily active users (DAU) measures the number of unique users who interact with Sky's vault system each day. This metric captures user engagement across Sky's lending and borrowing activities, categorizing users based on their specific actions: borrowing DAI/USDS, repaying debt, depositing collateral, withdrawing collateral, or trading through Peg Stability Modules (PSM). The calculation uses HyperLogLog approximation for efficient distinct counting and provides breakdowns across multiple dimensions including chain, user type, and product.

## 2. Data sources

- **Sky Vat contract function calls**: Parsed Ethereum function calls to Sky's core Vat contract, specifically the `frob` function which handles vault modifications including debt changes and collateral adjustments
- **Additional Sky contract functions**: `grab` and `fold` function calls used to identify and decode collateral types
- **Ethereum transaction records**: Transaction data providing user addresses and transaction metadata for linking function calls to actual users
- **Collateral type metadata**: Decoded collateral identifiers used to distinguish between regular vault operations and PSM trading activities

## 3. Step-by-step calculation

### Component A: Collateral type identification
- **Purpose**: Identify and decode Sky's collateral types to distinguish PSM operations from regular vault activities
- **Logic**: Extract collateral identifiers from Sky Vat contract `frob`, `grab`, and `fold` functions. Convert hex-encoded collateral IDs to readable names by removing hex prefixes and null bytes, then converting to UTF-8 strings. Create a reference mapping of collateral IDs to human-readable names
- **Key fields**: `i` (collateral ID), decoded `collateral_name` for filtering PSM operations

### Component B: User action classification
- **Purpose**: Classify user interactions based on vault modification patterns
- **Logic**: Parse `frob` function calls to examine debt changes (`dart`) and collateral changes (`dink`). Convert values from wei to standard units by dividing by 10^18. Classify users as 'Borrower' when debt increases (positive `dart`), 'Repayer' when debt decreases (negative `dart`), 'Depositor' when collateral increases (positive `dink`), 'Withdrawer' when collateral decreases (negative `dink`). For collateral types matching PSM patterns (names like 'PSM-%'), classify all users as 'Trader' regardless of direction. Filter to only successful transactions and exclude zero-value changes
- **Key fields**: `dart` (debt change), `dink` (collateral change), `i` (collateral type), transaction `status`

### Component C: User address resolution
- **Purpose**: Link anonymous contract function calls to actual user wallet addresses
- **Logic**: Join vault function calls with Ethereum transaction data using transaction hash. Extract the transaction initiator address (`from_address`) as the user performing the vault action. Apply incremental processing logic to only process recent data for efficiency
- **Key fields**: `transaction_hash` for joining, `from_address` as user identifier, `block_timestamp` for temporal filtering

### Component D: Daily user aggregation
- **Purpose**: Calculate daily distinct active users across different dimensions
- **Logic**: Truncate transaction timestamps to day boundaries. Group interactions by date, user type, product, and chain. Extract distinct user addresses for each combination. Set fixed values for market sector ('stablecoin_issuers'), business segment ('stablecoin_issuer'), product ('usds'), and chain ('ethereum')
- **Key fields**: `block_timestamp` truncated to day, `user_address` for distinct counting, `user_type` for segmentation

### Component E: HyperLogLog approximation
- **Purpose**: Enable efficient approximate distinct counting for large-scale aggregations
- **Logic**: Initialize HyperLogLog sketches with precision 15 for each daily user set. This allows approximate distinct counting while significantly reducing storage and computation requirements compared to exact counting. Group by timestamp, timestamp grain, chain, user type, and product ID
- **Key fields**: `user_dau` (user address) converted to HLL sketch, `hll` containing the approximation data structure

### Final aggregation
- **Combining components**: Create multiple dimensional views of the daily active user data by merging HLL sketches and extracting final counts. Generate separate breakdowns for project-level totals, chain-level totals, user-type-level totals, and product-level totals
- **Currency conversion**: No currency conversion required as this is a count-based metric
- **Time aggregation**: Data is aggregated at daily boundaries with month-level partitioning for efficient querying. Missing days are filled with zero values to ensure complete time series
- **Final calculation**: Extract final distinct user counts from merged HyperLogLog sketches using `hll_count.extract(hll_count.merge_partial(hll))`. Apply chain availability filtering and standardize output schema through the finalize_metric_user_dau macro

*This methodology was automatically generated by Token Terminal's methodology analyzer.*

Unique revenue-generating users over a fixed 24-hour window.

SkySKY

Active users (daily)

Active users (daily)

Competitive landscape

Chain breakdown