Materials

Content

Time Session URL
13:45 – 14:00 Embeddings Overview: what are they and why should I care? Slides
14:00 – 14:15 Interactive exploration of embeddings Imago Embedding Explorer
14:15 – 14:35 Semantic similarity [P] Notebook, JupyterLite
14:35 – 15:05 Exploring change over time with embeddings [P] Notebook, JupyterLite
15:05 – 15:45 Predictive modelling with embeddings [P] Notebook, JupyterLite

Data

To follow this workshop, you will need to download the following Imago’s data product:

Google Satellite Embedding V1 - London LSOAs (2020, 2024)

For this, follow these steps:

  1. Log in at Imago’s data catalogue. If you do not have an account, sign up for free with an email address1.
  2. Head to the data product’s page.
  3. You should now be able to see and download each year of embeddings at small area under the Resources section.
Tip

Once you are ready to graduate from our demo dataset, you can dive right into our newly released UK-wide data product of embeddings! You can access it at:

https://data.imago.ac.uk/datasets/google-satellite-embedding-v1-small-areas-2017-2024

This encompasses yearly embeddings from 2017 to 2024 at small area level (LSOA equivalent for all UK nations).

Footnotes

  1. The login is free and we only use it to keep track of usage, which our funder understandably wants to know about.↩︎