Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
Using Sentence Embeddings and Dimensionality Reduction to Visualize 26000 Dreams (josauder.github.io)
2 points by josauder__ on June 15, 2018 | hide | past | favorite | 1 comment


This started off as a project for my data visualization class, but it turned out really cool so I want to share it with everybody. This is an interactive visualization of over 26,000 English dreams, collected by the psychology department of UC Santa Cruz. The original data can be found [here](www.dreambank.net) - note that DreamBank is a completed project and they do not take in any more dream series.

The visualization is made with d3.js, the necessary data transformations are done in python, using some modern machine learning techniques - most notably sentence vectors, T-SNE, Autoencoders, and sentiment analysis. A full write-up of the techniques used can be found [here](http://josauder.github.io/dreambank_visualized/dreambank_vis...), and the code (which is very messy, please don't judge me) can be found [here](http://github.com/josauder/dreambank_visualized).

I hope you like it!

Feel free to follow me on [Twitter](http://twitter.com/josauder), I am rather new but there is more content to come!




Consider applying for YC's Fall 2025 batch! Applications are open till Aug 4

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: