How to Use Redis to Count Website Visits: Hash, Bitset, and Probabilistic Algorithms

Using Hash

Redis hash is a basic data structure where each key maps to a hash table; on collisions a linked list is used.

When a user visits, use their user ID (or a randomly generated key for anonymous users) as the field and a composite key of URI and date as the Redis key; the HSET command stores a value of 1.

To get the daily visit count for a page, the HLEN command returns the number of fields, i.e., unique visitors.

Advantages: Simple, easy to implement, fast queries, high data accuracy.

Disadvantages: High memory usage; performance degrades as the number of keys grows, making it unsuitable for billions of page views.

Using Bitset

By representing each user ID as a single bit in a bitmap, memory usage can be reduced dramatically; for example, 100 million users require only about 12 MB.

Redis provides the SETBIT command to mark a user as having visited a page and GETBIT to check the status; BITCOUNT aggregates the total number of set bits for the day.

Advantages: Very low memory consumption, fast queries, can target individual users.

Disadvantages: May consume more memory than the hash method when user IDs are sparse, and requires a mapping for anonymous users.

Using Probabilistic Algorithm

For extremely large traffic where exact counts are unnecessary, Redis’s HyperLogLog ( PFADD / PFCOUNT ) offers cardinality estimation with about 0.81 % error and uses roughly 12 KB per key.

Advantages: Minimal memory footprint, suitable for massive user bases.

Disadvantages: Cannot retrieve individual user information and introduces a small counting error.

These three methods provide trade‑offs between accuracy, memory usage, and scalability for counting website visits.