Imagine having a bird’s-eye view of the entire landscape of cancer research. That’s exactly what a fascinating project has achieved, mapping out the connections and patterns within 2.5 million cancer research papers from 2010 to 2025.
The creator of this project used a transformer neural network to embed the titles and abstracts of these papers, then projected them with UMAP and clustered them with Leiden. The result is an interactive atlas that reveals 46 distinct communities, each representing a core area of focus within the field.
You can explore the atlas in full on the creator’s website, where you can zoom in on specific areas and click on paper titles to read them. The clusters span the breadth of cancer research, including cancer types, treatment strategies, molecular and cellular biology, clinical and diagnostic domains, and cross-cutting and emerging themes.
This visual representation of cancer research is not only stunning but also reveals hidden patterns and connections that can inform future research. It’s a testament to the power of data visualization in advancing our understanding of complex fields like cancer research.
## Exploring the Atlas
The atlas is divided into clusters that cover various aspects of cancer research. You’ll find clusters dedicated to specific cancer types, such as breast, lung, and pancreatic cancer. There are also clusters focused on treatment strategies, like immunotherapy and targeted therapies.
Other clusters delve into molecular and cellular biology, exploring topics like signaling pathways, non-coding RNAs, and epigenetics. You’ll also find clusters centered on clinical and diagnostic domains, including patient outcomes, imaging, and diagnostics.
Perhaps most intriguing are the cross-cutting and emerging themes, which highlight areas like the tumor microenvironment, inflammation, and AI in oncology.
## The Power of Data Visualization
This project demonstrates the power of data visualization in uncovering hidden patterns and connections within complex data sets. By mapping the landscape of cancer research, researchers and clinicians can identify areas of overlap, spot trends, and pinpoint potential areas for further exploration.
The atlas is not only a valuable resource for the cancer research community but also a testament to the potential of data visualization to drive innovation and progress in various fields.
—
*Explore the cancer atlas: