Validators process transactions and receive fees from users (supply-side fees).

Aptos is a smart contract platform for decentralized applications (a blockchain). Aptos was founded by Mo Shaikh and Avery Ching and launched in 2022.

APT token holders own and govern the protocol. They currently don’t take a cut of the total transaction fees paid by users (revenue). Aptos’ investors include Binance Labs, Jump Crypto, Apollo, etc.

governance

github

twitter

discord

website

Blockchains (L1)

Aptos

# Aptos Active addresses (weekly) Methodology

## 1. Overview

Active addresses weekly measures the number of unique addresses that initiated at least one transaction on the Aptos blockchain within any 7-day rolling window. This metric answers the business question "How many distinct users are actively transacting on Aptos over weekly periods?" The calculation uses HyperLogLog probabilistic counting algorithms to efficiently process large volumes of transaction data while maintaining accuracy for unique address counts across overlapping time windows.

## 2. Data sources

The calculation relies on these core data sources:
- **Aptos transaction records**: Complete transaction history from the Aptos blockchain including transaction hashes, sender addresses, timestamps, and gas usage data
- **Transaction metadata**: Block timestamps for time-based aggregation, sender addresses for unique user identification, and gas pricing information
- **Genesis data handling**: Hardcoded genesis timestamp of October 12th, 2022 to ensure consistent historical baseline for the Aptos mainnet launch

## 3. Step-by-step calculation

### Component A: Transaction data preparation
- **Purpose**: Extract and standardize raw Aptos transaction data with proper time boundaries
- **Logic**: All transactions from Aptos blockchain data are collected, with gas unit prices normalized by dividing by 10^8 to convert from octas to APT. A special handling mechanism ensures genesis transactions (those with timestamps before 2000-01-01) are assigned the hardcoded mainnet launch timestamp of October 12th, 2022, while regular transactions maintain their original block timestamps. This creates a clean dataset with consistent temporal boundaries.
- **Key fields**: `block_timestamp` for time-based grouping, `sender` addresses as the unique entities to count, `transaction_hash` for transaction identification, `gas_used` and `gas_unit_price` for transaction metadata

### Component B: Daily active user calculation  
- **Purpose**: Identify unique active addresses on a daily basis
- **Logic**: Transaction data is grouped by day using timestamp truncation to day granularity. For each day, sender addresses are deduplicated to create a list of unique active addresses. Each record represents one unique address that was active on a specific date, along with classification metadata including chain identifier, user type as "Transaction sender", market sector as "blockchains_l1", and business segment as "blockchain".
- **Key fields**: `timestamp` truncated to day level, `user_dau` containing unique sender addresses, `chain` set to "aptos", `user_type` labeled as "Transaction sender"

### Component C: HyperLogLog sketch creation
- **Purpose**: Convert daily active addresses into probabilistic counting data structures for efficient weekly aggregation
- **Logic**: Daily active user data is transformed into HyperLogLog sketches using precision parameter 15. Each day's unique addresses are processed through the HLL initialization function, creating compact probabilistic representations that enable efficient counting of unique elements across large datasets. The sketches are grouped by timestamp, timestamp grain, chain identifier, user type, and product identifier.
- **Key fields**: `hll` sketches containing probabilistic representations of daily unique addresses, `timestamp` for temporal grouping, `chain_id` for blockchain identification

### Component D: Rolling window generation
- **Purpose**: Create overlapping 7-day windows for weekly active user calculation
- **Logic**: A date spine is generated covering all days from the minimum timestamp in the daily data through yesterday. For each date in this spine, a 7-day window is created by generating timestamps from 6 days prior through the current date. This creates overlapping windows where each date represents the end of a 7-day period, ensuring continuous weekly metrics without gaps.
- **Key fields**: `timestamp` as window end dates, `joinday` representing each day within the 7-day windows

### Component E: Weekly aggregation with HyperLogLog
- **Purpose**: Calculate unique addresses across each 7-day rolling window
- **Logic**: The rolling windows are joined with daily HLL sketches, matching each day within a 7-day window to its corresponding HLL data. For each 7-day window, the HLL sketches are merged using partial merge operations, then the final unique count is extracted. This probabilistic approach efficiently handles the complex set operations required for counting unique addresses across overlapping time periods while maintaining high accuracy.
- **Key fields**: `user_wau` containing the count of unique addresses active within each 7-day window, `timestamp` marking the end of each weekly period

### Component F: Dimensional breakdowns
- **Purpose**: Provide weekly active addresses across different analytical dimensions
- **Logic**: The weekly active user data is broken down into four dimensional views: chain-level aggregation showing overall Aptos activity, project tag aggregation for protocol-level analysis, user type breakdown for different address categories, and product-level segmentation. Each breakdown uses the same HLL merge and extract operations but groups by different dimensional attributes, creating parallel analytical views of the same underlying weekly activity data.
- **Key fields**: `dimensions` indicating the breakdown type, `user_wau` with dimension-specific unique address counts, dimension-specific identifiers (`chain_id`, `product_id`, `user_type`)

### Final aggregation
- **Combining components**: Weekly active user data from all dimensional breakdowns is combined using union operations, creating a comprehensive dataset with chain and project tag dimensions (other dimensions are filtered out in the final model)
- **Time aggregation**: The metric maintains daily resolution with each record representing unique addresses active in the 7-day period ending on that date
- **Final calculation**: The `finalize_metric_active_addresses_weekly` macro applies schema standardization, ensures proper data types, fills any missing time periods with zero values, and applies the final column aliasing from `user_wau` to `active_addresses_weekly` to match the standard metric naming convention

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

Unique active addresses over a rolling 7-day window.

AptosAPT

Active addresses (weekly)

Active addresses (weekly)

Competitive landscape