I used camelAI with a ClickHouse database of every HN story to do all analysis. You can use it for free with no login here to explore the data interactively yourself.
Hacker News is a real-time barometer of developer excitement. I mined every story title from February 2007 to June 2025 and asked three questions:
The chart below shows monthly headline counts for the eight most-talked-about engines, plus DuckDB (small base, big growth).
Click on the individual database names to toggle their visibility. Double click to isolate a single database.
PostgreSQL’s curve is a near-monotonic climb; by 2020 it dwarfs every other line. MySQL dominates the pre-2012 era, then flat-lines. MongoDB peaks around 2013, then slips as SQL engines add JSON support. ClickHouse (2016) and DuckDB (2020) rocket up in the analytics renaissance. Redis & SQLite are steady, underscoring their “invisible infrastructure” roles.
I compared the last 12 months (Jul 2024 – Jun 2025) with the previous 12 months, then fit a linear regression over the last 36 months to capture trajectory.
Engine | Headlines 2024-25 | 2023-24 | YoY Δ % | 3-yr slope (mentions / mo) |
---|---|---|---|---|
DuckDB | 225 | 149 | +50.7 % | +0.44 |
ClickHouse | 211 | 170 | +24.1 % | +0.39 |
Supabase | 105 | 84 | +25.0 % | +0.12 |
PostgreSQL | 1 229 | 1 315 | −6.6 % | +1.22 |
Snowflake | 106 | 156 | −32.1 % | +0.01 |
Redis | 180 | 236 | −23.7 % | +0.07 |
MySQL | 164 | 185 | −11.4 % | +0.03 |
MongoDB | 96 | 78 | +23.1 % | +0.03 |
SQLite | 436 | 407 | +7.1 % | +0.04 |
DynamoDB | 25 | 38 | −34.2 % | −0.08 |
BigQuery | 20 | 35 | −42.9 % | −0.05 |
Redshift | 10 | 18 | −44.4 % | −0.05 |
TimescaleDB | 3 | 4 | −25.0 % | −0.01 |
DuckDB tops the growth chart: half of its lifetime headlines appeared this year. ClickHouse maintains double-digit gains on a larger base, buoyed by vector search & managed services. Supabase benefits from the “Firebase, but OSS” narrative. PostgreSQL’s YoY dip merely reflects an all-time-high 2023; its 3-yr slope (+1.22) is the steepest of all. Cloud-native SaaS engines (DynamoDB, BigQuery, Redshift) are down sharply in share of conversation.
Headline counts aren’t everything, but they do reveal mind-share:
Engine | Peak year | Peak headlines | 2024-25 headlines | % off peak |
---|---|---|---|---|
MongoDB | 2013 | 242 | 96 | −60 % |
MySQL | 2009 | 267 | 164 | −39 % |
DynamoDB | 2019 | 59 | 25 | −58 % |
BigQuery | 2020 | 62 | 20 | −68 % |
Redshift | 2016 | 61 | 10 | −84 % |
Why? Maturity – less novelty, fewer “Show HN” launches. OSS competition – Postgres extensions + parquet/iceberg lakehouses replace single-purpose stores. Discussion shift – Cost, lock-in, and serverless topics often omit the product name.
(July 2024 – June 2025)
Beyond raw headline counts, weighting stories by Hacker News points and comments reveals how intensely developers engage with each database.
Database | Stories | Total Points | Total Comments | Avg Points per Story | Avg Comments per Story |
---|---|---|---|---|---|
PostgreSQL | 1,229 | 26,185 | 8,666 | 21.3 | 7.1 |
SQLite | 436 | 17,598 | 5,136 | 40.4 | 11.8 |
Redis | 180 | 5,995 | 2,392 | 33.3 | 13.3 |
DuckDB | 225 | 5,001 | 1,140 | 22.2 | 5.1 |
ClickHouse | 211 | 3,147 | 891 | 14.9 | 4.2 |
Snowflake | 106 | 1,212 | 518 | 11.4 | 4.9 |
MySQL | 164 | 1,167 | 514 | 7.1 | 3.1 |
Supabase | 105 | 849 | 548 | 8.1 | 5.2 |
MongoDB | 96 | 572 | 362 | 6.0 | 3.8 |
BigQuery | 20 | 342 | 142 | 17.1 | 7.1 |
DynamoDB | 25 | 183 | 85 | 7.3 | 3.4 |
Redshift | 10 | 37 | 1 | 3.7 | 0.1 |
TimescaleDB | 3 | 3 | 1 | 1.0 | 0.3 |
Aggregate story count, points, and comments for each database in the last 12 months.
SQLite punches far above its weight:
Only one-third the story count of Postgres but two-thirds the total points.
Raw mention counts reveal frequency; weighting by points and comments captures intensity of community reaction.
Taken together, frequency and engagement metrics offer a richer, more nuanced perspective on developer sentiment.
Both raw popularity and weighted engagement metrics show clear trends:
Questions you can run next:
Run any query on the same dataset here, no login required.