OpenChat users can send messages to each other containing tokens such as ICP. Users can upgrade to Diamond membership to have access to extra or enhanced features.

OpenChat is a chat application running end-to-end on the Internet Computer blockchain.

CHAT tokenholders own and govern the protocol.

github

twitter

website

Social

OpenChat

# OpenChat Active users (weekly) Methodology

## 1. Overview

Weekly Active Users (WAU) measures the number of unique users who were active on the OpenChat platform within any 7-day rolling window. This metric provides insight into user engagement and platform growth by tracking how many distinct users interact with OpenChat's social features over weekly periods. The calculation uses HyperLogLog sketches for efficient approximate cardinality estimation across overlapping time windows, enabling analysis across multiple dimensional breakdowns including overall project metrics, chain-level analysis, user type segmentation, and product-specific engagement.

## 2. Data sources

The calculation relies on the following data sources:
- **OpenChat event stream**: User activity events from OpenChat's `events_v4` data stream containing timestamped user interactions
- **User online events**: Specific event type indicating when users are active on the OpenChat platform 
- **Internet Computer metadata**: Blockchain context information for the Internet Computer network where OpenChat operates
- **Platform metadata**: Social media sector classification and business segment categorization

## 3. Step-by-step calculation

### Component A: Event processing and deduplication
- **Purpose**: Extract and clean raw user activity events from OpenChat's event stream
- **Logic**: Raw events are filtered to include only `user_online` events occurring since 2024-02-14. Events are deduplicated using a combination of `timestamp`, `user`, `name`, `source`, and JSON-serialized `payload` fields to ensure each unique user activity is counted only once. A row number is assigned to each event group and only the first occurrence is retained.
- **Key fields**: `timestamp` (event time), `name` (event type filter), `user` (user identifier), `source` (event origin), `payload` (event data)

### Component B: User identification and daily aggregation  
- **Purpose**: Identify unique users active each day on the OpenChat platform
- **Logic**: Processed events are grouped by day using date truncation of the `timestamp` field. Each unique `user` identifier active within a given day contributes to that day's active user count. Additional metadata is applied including chain classification as 'internetcomputer', market sector as 'social', and business segment as 'social'. Product-specific and user type classifications are left as null for this general platform metric.
- **Key fields**: `timestamp` (truncated to day), `user_address` (unique user identifier), `chain` (blockchain network), `market_sector_id` and `business_segment_id` (categorization)

### Component C: HyperLogLog sketch initialization
- **Purpose**: Convert daily user sets into probabilistic data structures for efficient weekly calculations
- **Logic**: Daily active user counts are converted into HyperLogLog sketches using `hll_count.init()` with precision parameter 15. This enables approximate distinct counting when merging overlapping time windows. Users are grouped by `timestamp`, `timestamp_grain`, `chain_id`, `user_type`, and `product_id` dimensions to maintain granular breakdowns while creating the probabilistic sketches.
- **Key fields**: `hll` (HyperLogLog sketch), dimensional grouping fields for later aggregation

### Component D: Date spine and rolling window creation
- **Purpose**: Generate continuous time series and create 7-day rolling windows for weekly calculations
- **Logic**: A date spine is generated from the minimum date in the daily user data through yesterday using `generate_timestamp_array()` with daily intervals. For each date in the spine, a 7-day window is created by generating timestamps from 6 days prior through the current day. This ensures complete 7-day windows are available for every date and handles missing days in the source data.
- **Key fields**: `timestamp` (window end date), `joinday` (individual days within the 7-day window)

### Component E: Weekly active user calculation
- **Purpose**: Calculate weekly active users by merging daily HyperLogLog sketches across 7-day windows
- **Logic**: For each day's 7-day window, the corresponding daily HyperLogLog sketches are collected and merged using `hll_count.merge_partial()`. The merged sketch is then converted to a cardinality estimate using `hll_count.extract()` to produce the weekly active user count. This process accounts for users who may be active on multiple days within the week by counting them only once.
- **Key fields**: `user_wau` (weekly active user count), dimensional fields maintained from daily processing

### Component F: Dimensional breakdowns
- **Purpose**: Create multiple analytical views of weekly active users across different dimensions
- **Logic**: Weekly active user calculations are performed across four dimensional breakdowns: overall project level (project_dbt_tag), chain level, user type level, and product level. Each breakdown groups and aggregates the HyperLogLog sketches appropriately while setting non-relevant dimension fields to null. The results from all dimensional breakdowns are combined using union operations to create a comprehensive dataset.
- **Key fields**: `dimensions` (breakdown type), `chain_id`, `user_type`, `product_id` (dimensional values), `user_wau` (metric value)

### Final aggregation
- **Combining components**: All dimensional breakdowns are combined through union operations to create a unified weekly active user dataset with multiple analytical perspectives
- **Currency conversion**: Not applicable for this user count metric
- **Time aggregation**: Daily boundaries are maintained with 7-day rolling windows, ensuring each day has a corresponding weekly active user calculation
- **Final calculation**: The final weekly active user count represents the approximate number of distinct users who were active within any 7-day period ending on each date, calculated using HyperLogLog cardinality estimation to efficiently handle the overlapping nature of rolling windows

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

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

OpenChatCHAT

Active users (weekly)

Active users (weekly)

Competitive landscape