Learn AI · Level 2 · How AI works · step 6 of 19 · word 4 of 4
What is a data lake?
A data lake is a large, low-cost store that keeps raw data of every kind, from tables to images, text and logs, as it arrives, to be sorted out later when someone needs it.
In plain English
If a data warehouse is a library, a data lake is a huge storeroom. Everything goes in as it is: spreadsheets, photos, chat logs, sensor readings. Nothing has to be cleaned first, which makes it cheap and flexible. That suits AI, because training data is often messy text, images and audio.
An example
A car maker pours years of camera footage from its test cars into a data lake. Later, a team pulls out every clip filmed in rain to train a driver-assistance model.
Watch out
Without labels and rules, a lake turns into a “data swamp”: nobody knows what is in it, how old it is or whether it may be used. A newer design, the lakehouse, adds warehouse-style order on top of a lake.
Learn AI path · Level 2 · How AI works · step 6 of 19
Done? Next stepClassic models →See all of level 2