website

twitter

Exchanges (DEX)

pump.fun

# pump.fun Active users (weekly) Methodology

## 1. Overview

Weekly Active Users (WAU) measures the number of unique users who have traded on Pump.fun within any 7-day rolling window. This metric helps assess user engagement and platform activity trends for the Pump.fun decentralized exchange on Solana. The calculation uses HyperLogLog probabilistic counting to efficiently process large volumes of user data and provides multiple dimensional breakdowns including project-level, chain-level, user type, and product-level views.

## 2. Data sources

The calculation relies on the following raw data sources:

- **Solana DEX trade events**: Raw transaction data from Solana's `solana_dex_trades_v2` dataset, specifically filtered for trades executed through the Pump.fun program address `6EF8rrecthR5Dkzon8Nwu78hRvfCKubJ14M5uBEwF6P`
- **Token metadata**: Solana token information including decimal places for proper amount normalization
- **Transaction signers**: Wallet addresses that initiated the trading transactions, used as unique user identifiers

## 3. Step-by-step calculation

### Component A: Trade data extraction and processing
- **Purpose**: Extract and normalize Pump.fun trading activity from raw Solana transaction data
- **Logic**: The system filters Solana DEX trades to include only transactions executed through the specific Pump.fun program address. For each trade, it captures the `transaction_id`, `block_timestamp`, `transaction_signer` (wallet address), base and quote token addresses, amounts, and instruction type. Raw token amounts are normalized by dividing by 10 raised to the power of each token's decimal places using token metadata lookups. The absolute values of amounts are taken to ensure consistent positive values regardless of trade direction.
- **Key fields**: `transaction_signer` (user wallet address), `block_timestamp` (transaction timing), `base_amount` and `quote_amount` (normalized token quantities), `instruction_type` (trade operation), `product_id` (set to 'pump.fun')

### Component B: Daily active user calculation
- **Purpose**: Identify unique users who traded each day on Pump.fun
- **Logic**: The system processes the normalized trade data to identify daily unique users by truncating the `block_timestamp` to day granularity and selecting distinct `transaction_signer` values for each day. All users are classified as 'Trader' user type since the metric focuses on trading activity. The output includes the timestamp (daily), chain ('solana'), user type ('Trader'), product identifier ('pump.fun'), and the unique user address.
- **Key fields**: `timestamp` (daily truncated), `user_dau` (unique signer address), `user_type` (always 'Trader'), `chain` ('solana')

### Component C: HyperLogLog conversion and windowing preparation
- **Purpose**: Convert daily user data to probabilistic counting format and prepare for rolling window analysis
- **Logic**: Daily active user data is converted to HyperLogLog (HLL) format using `hll_count.init()` with precision 15 for efficient unique counting operations. The system creates a complete date spine from the minimum date in the DAU data to the current date minus one day, ensuring no missing timestamps in the time series. For each date in the spine, it generates a 7-day lookback window by creating join days from 6 days prior through the current date using `generate_timestamp_array()`.
- **Key fields**: `hll` (HyperLogLog encoded user data), `timestamp` (daily dates), `joinday` (dates within 7-day window)

### Component D: Weekly active user window aggregation
- **Purpose**: Calculate unique users within each 7-day rolling window using HLL operations
- **Logic**: The system joins the date spine with HLL-encoded daily user data where the join days match the DAU timestamps. This creates overlapping 7-day windows for each date. Within each window, HLL data structures are merged using `hll_count.merge_partial()` to combine unique user counts across the 7 days, then extracted using `hll_count.extract()` to get the final unique user count. This approach efficiently handles the probabilistic counting of unique users across time periods.
- **Key fields**: `timestamp` (window end date), `hll` (merged HLL data), dimensions identifying chain, user type, and product

### Final aggregation
- **Combining components**: The system creates four separate dimensional breakdowns: project-level (all dimensions nulled), chain-level (grouped by `chain_id`), user type-level (grouped by `user_type`), and product-level (grouped by `product_id`). Each breakdown aggregates the HLL data appropriately and sets a `dimensions` field to identify the aggregation level. All breakdowns are combined using UNION ALL operations.
- **Currency conversion**: Not applicable as this metric counts users rather than monetary values
- **Time aggregation**: Uses rolling 7-day windows where each day represents the unique users who were active in that day plus the previous 6 days
- **Final calculation**: The `finalize_metric_user_wau` macro applies standardized formatting, sets the data identifier to 'pump.fun', fills missing days with zeros using the fill_zeros strategy, and ensures proper schema alignment. The final output provides daily WAU values with multiple dimensional breakdowns for comprehensive user activity analysis.

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

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

pump.funPUMP

Active users (weekly)

Active users (weekly)

Competitive landscape