Places to get data from

2024 February 26

Over the past year, I have acumulated a list of different places where you can get high quality datasets. This is the list in no particular order on where to get data for different type of problems.

• Dataset Research Engine by Google: You might never have heard of it, but Google has a search engine specifically for datasets. The results vary, but it can be a good starting point.

• huggingface: A hub from anything from models to super intricate datasets. In general, I use this more for Deep Learning than for ML.

• kaggle: It is a community / competition playground from Google. Some of the datasets are really high quality but it can be really mixed. They have some high quality ML datasets.

• paperswithcode: In here, there are a lot of what could be considered the classic datasets. The datasets that you normally use to demonstrate State of the Art on a new method. It is more academic. Normally it is high quality.

• Data in Brief: This is way more academic. It is an open access journal that is about open data.

• Data world: It feels a bit spammy, and the results can be mixed.

• Health Data: Health data from the US.

Last updated: 2024-02-26