Thread by @findingdesti

Finally took the time to figure out what the duck is #DuckDB and why I should care.

There’s a lot of technical jargon that gets thrown and I wish there was a plain English explanation of why analysts should be excited.

Here’s my attempt at doing exactly that (1/🦆)

#DuckDB brands itself as SQLite for analytics. If you’re like me you’ve probably heard of SQLite somewhat superficially.

Little did I know that SQLite was a low-key game-changing piece of software.

SQLite removes the complexity of client server databases (think Oracle, SQL Server) and instead stores the database in a single file. This file is stored locally on your disk which makes reads and writes very fast, and the database extremely portable and lightweight.

The success of this architecture made SQLite the go-to choice for an application that needed a database to store information. In this day and age that’s pretty much every application - web browsers, smart phones, laptops, and even thermostats.

This has made SQLite one of the most ubiquitous pieces of software ever with over 2 billion deployments!

The key to this success was just how easy it was for applications to use SQLite to read/write information - think OLTP queries.

#DuckDB is the OLAP analog to this. Setting it up is super easy and all you need is DBeaver and a CSV file that you want to query. You essentially write SQL directly on the file making it so much more frictionless to use.

You don’t need to know pandas or spin up a database which means that #DuckDB is poised to become the very first exposure to SQL for beginners. This is a great position to have.

#DuckDB SQL is also extremely feature rich and has a lot of handy functions like EXCEPT and REPLACE. Read more about this on their blog.

However, this is just the tip of the iceberg. If SQLite is anything to go by, there’s no reason why #DuckDB doesn’t get widely adopted into applications.

This would basically mean that every application around us gets shipped with a native data warehouse capable of answering complex queries.

In financial apps this could show you way more that how much you spent over a month. Now you can see average spend per day based on your location or seasonality.

In freemium apps this could mean nudging users in real time to upgrade based on usage and product signals rather than batch analytics followed by an email blast from marketing. True real time analytics is possible.

The possibilities are endless and I’m really excited to see where @duckdb falls in the data stacks of the present and future.