Arbitrum One batches users' transactions and submits them to Ethereum, paying a share of the transaction fees to Ethereum validators (expense).

Arbitrum One is an L2 smart contract platform for decentralized applications on Ethereum. Arbitrum was founded by Offchain Labs and launched in 2021.

ARB tokenholders own and govern the protocol. They currently run the only sequencer and take a cut of the total transaction fees paid by users (revenue). Arbitrum’s investors include Polychain, Alameda Research, Coinbase Ventures, etc.

github

twitter

discord

etherscan

website

Blockchains (L2)

Arbitrum One

# Arbitrum One Active users (weekly) Methodology

## 1. Overview

Weekly Active Users (WAU) measures the number of unique users who performed at least one transaction on the Arbitrum network within any 7-day rolling window. This metric provides insight into user engagement and activity patterns by tracking how many distinct addresses sent transactions over weekly periods. The calculation uses HyperLogLog algorithms for efficient approximate counting of large user sets and creates multiple dimensional breakdowns to analyze user activity across different segments.

## 2. Data sources

- **Arbitrum transaction records**: Complete transaction data from the Arbitrum blockchain, including transaction timestamps, sender addresses (`from_address`), recipient addresses (`to_address`), gas usage, and gas prices
- **Token metadata**: Token information including the WETH contract address (`0x82af49447d8a07e3bd95bd0d56f35241523fbab1`) and token symbols for gas fee calculations
- **Chain availability data**: Network availability filters to ensure data quality and exclude periods when chain data may be incomplete

## 3. Step-by-step calculation

### Component A: Daily active user extraction
- **Purpose**: Identify unique users (transaction senders) active each day on Arbitrum
- **Logic**: Extract all Arbitrum blockchain transactions and identify distinct sender addresses (`from_address`) for each day. Transactions are grouped by truncating `block_timestamp` to day granularity. Each unique sender address represents one daily active user. Additional metadata is added including chain identifier, user type classification as "Transaction sender", market sector as "blockchains_l2", and business segment as "blockchain"
- **Key fields**: `timestamp` (day-truncated), `chain`, `user_type`, `product_id`, and `user_dau` (the sender address)

### Component B: HyperLogLog initialization
- **Purpose**: Convert daily user counts into HyperLogLog sketches for efficient approximate counting
- **Logic**: Transform the daily active user data by initializing HyperLogLog sketches using `hll_count.init()` with precision parameter 15. This creates compact probabilistic data structures that can efficiently estimate distinct user counts and be merged across time periods. Each daily user set becomes an HLL sketch grouped by timestamp, chain, user type, and product dimensions
- **Key fields**: `hll` (HyperLogLog sketch), `timestamp`, `chain_id`, `user_type`, `product_id`

### Component C: Date spine and window generation  
- **Purpose**: Create a complete date range and generate 7-day rolling windows for WAU calculation
- **Logic**: Build a datespine covering all dates from the earliest DAU data to current date using `generate_timestamp_array()`. For each date in the spine, create a 7-day window by generating timestamps for the current date plus the previous 6 days. This ensures every date has a complete 7-day lookback window even when some days have no user activity. The window spans from 6 days before each target date through the target date itself
- **Key fields**: `timestamp` (target date), `joinday` (each day in the 7-day window)

### Component D: Window user aggregation
- **Purpose**: Combine HLL sketches across 7-day windows to calculate weekly active users  
- **Logic**: Join the date windows with the daily HLL data, matching each `joinday` in the 7-day window to corresponding dates in the daily HLL sketches. This creates a dataset where each target date is associated with all HLL sketches from its 7-day window. The HLL sketches are then merged using `hll_count.merge_partial()` to combine user counts across the window period, and finally extracted using `hll_count.extract()` to produce the approximate weekly active user count
- **Key fields**: `user_wau` (extracted weekly active user count), maintaining all dimensional attributes

### Component E: Dimensional breakdowns
- **Purpose**: Generate WAU metrics across multiple analytical dimensions
- **Logic**: Create four separate dimensional views of the WAU data. Project-level breakdown aggregates all users across the entire Arbitrum ecosystem. Chain-level breakdown maintains chain granularity. User-type breakdown segments by transaction sender classification. Product-level breakdown segments by specific product identifiers. Each breakdown uses appropriate grouping and sets unused dimension fields to null. All breakdowns are then combined using union operations
- **Key fields**: `dimensions` (breakdown type), `chain_id`, `product_id`, `user_type`, `user_wau`

### Final aggregation
- **Combining components**: The multiple dimensional breakdowns are unioned together to create a comprehensive dataset with WAU metrics across all analytical dimensions
- **Time aggregation**: Daily boundary logic using day-truncated timestamps with 7-day rolling windows calculated for each date
- **Final calculation**: Weekly active users represent the approximate count of distinct addresses that sent at least one transaction within any 7-day period ending on each date, calculated using HyperLogLog merge and extract operations for scalable approximate counting
- **Data quality**: Incremental processing with safety margins and lookback periods, chain availability filtering, and zero-filling for missing dates using the `finalize_metric_user_wau` macro

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

Unique revenue-generating users over a rolling 7-day window.

Arbitrum OneARB

Active users (weekly)

Active users (weekly)

Competitive landscape

Chain breakdown