Live data from Hacker News

Hacker News Data Map [180MB]

lmcinnes.github.io

1–10 of 27 posts

Re: Hacker News Data Map [180MB]

#3
Maybe add [180MB] to the title, similar to how videos or pdfs are tagged? It starts loading that immediately when you open the page, which would be 18% of my data bundle if I had been on mobile

(This is actually transferred bytes btw, based on seeing ~12MiB/s for ~15 seconds in the system monitor)

Edit: some people are saying they can't view it, especially on mobile browsers. Here's some screenshots:

- Landing overview https://snipboard.io/YTQRZc.jpg

- Zooming into the center, hovering over an item that is too small to see but the title shows in a tooltip: https://snipboard.io/xOvA47.jpg

- Zoomed in further still, now an individual item can be targeted easily and there are lines delimiting topics (looking like height lines on a map): https://snipboard.io/P6UVAv.jpg

- Hovering over the year selector on the bottom left, same zoom position for comparison: https://snipboard.io/VDW2JI.jpg

Clicking the year seems not to do anything, you can't lock into that view. Clicking a title opens the page, not the discussion thread.

---

Looking into the corresponding GitHub repository (I wonder if they have a bandwidth limit for repositories or if it will foot any bill), https://github.com/lmcinnes/datamapplot_examples>, there's also a visualization for Wikipedia which is a bit less heavy: https://lmcinnes.github.io/datamapplot_examples/Wikipedia_da... (screenshot https://snipboard.io/M9GRQt.jpg>)

Re: Hacker News Data Map [180MB]

#5
post #3

Maybe add [180MB] to the title, similar to how videos or pdfs are tagged? It starts loading that immediately when you open the page, which would be 18% of my data bundle if I had been on mobile (This is actually transferred bytes btw, based on seeing ~12MiB/s for ~15 seconds in the system monitor) Edit: some people are saying they can't view it, especially on mobile browsers. Here's some screenshots: - Landing overvi…

180MB download per HN visitor isn't going to be fun for the server either.

More civilized would be a photo snapshot + optional link to 180MB download for interactive UX.

Re: Hacker News Data Map [180MB]

#7
post #3

Maybe add [180MB] to the title, similar to how videos or pdfs are tagged? It starts loading that immediately when you open the page, which would be 18% of my data bundle if I had been on mobile (This is actually transferred bytes btw, based on seeing ~12MiB/s for ~15 seconds in the system monitor) Edit: some people are saying they can't view it, especially on mobile browsers. Here's some screenshots: - Landing overvi…

180MB download per HN visitor isn't going to be fun for the server either. More civilized would be a photo snapshot + optional link to 180MB download for interactive UX.

A photo could work as a quick preview indeed. As another idea for large content spiking in popularity, another solution may be something like webtorrent, or whatever peertube uses

Or a vector map, loading data as needed for the region you're zooming into

Re: Hacker News Data Map [180MB]

#8
post #3

Maybe add [180MB] to the title, similar to how videos or pdfs are tagged? It starts loading that immediately when you open the page, which would be 18% of my data bundle if I had been on mobile (This is actually transferred bytes btw, based on seeing ~12MiB/s for ~15 seconds in the system monitor) Edit: some people are saying they can't view it, especially on mobile browsers. Here's some screenshots: - Landing overvi…

180MB download per HN visitor isn't going to be fun for the server either. More civilized would be a photo snapshot + optional link to 180MB download for interactive UX.

180MB is probably fine for most servers (especially CDNs), to be honest. My M1 MacBook with 16 gigabytes of RAM is struggling to load/display the data though.

Re: Hacker News Data Map [180MB]

#10
I like how Web Development and User Experience grouping is way outside the central bubble.

Nonetheless, great visualization of a lot of data. I need to learn more about this:

UMAP: https://umap-learn.readthedocs.io/en/latest/

Nomic-Embed: https://www.nomic.ai/blog/posts/nomic-embed-text-v1

The visual groupings aren't perfect. For example, there are a quite a few COVID-19 tagged articles before 2020.

Post reply on HN