Excited to share our design concept for embedding temporal data visualizations into urban streets @EuroVis next week.
| Lab | https://www.hdilab.org |
Till Nagel
- 358 Followers
- 156 Following
- 52 Posts
Designing, developing, researching and teaching visualizations. Professor for visual analytics at Hochschule Mannheim. https://www.hdilab.org
HeatHub is a visualization system to show spatial, hierarchical and temporal dimensions of heat meter data for the former Berlin airport. Developed by students of the Mannheim University of Applied Sciences in cooperation with the Urban Tech Republic, Berlin TXL. https://www.hdilab.org/projects/heathub/
We had the pleasure to not only present and discuss the prototype but visit the former airport facilities. That was awesome!
This week, students showcased their heat meter analysis dashboard at Urban Tech Republic, Berlin. After weeks of workshops and visualization design, the talented team impressed with their data-driven prototype. More info soon.
Final touches underway for MIX Day 2023! Getting ready for an interdisciplinary, multi-space escape room experience open to all students. (including data vis puzzles! 📊🕵️♂️)
New podcast on Making with Data! Check out the newest PolicyViz episode with Samuel Huron, myself, Lora Oehlberg and Wesley Willet chatting with host Jonathan Schwabish about data physicalizations and the fascinating projects from the book.
https://policyviz.com/podcast/episode-245-making-with-data/
https://policyviz.com/podcast/episode-245-making-with-data/
All screenshots are the beginning of a chat. Occasionally, Bing responds that it cannot see the image but does describe it nevertheless. It hallucinates occasionally, e.g. identifying China in the chocolate/Nobel prize chart.
For rendered graphs it seems to work better, though it's not fully clear what is based on the image and what on its corpus (or search results) if depicting well known example data sets.
For complex sketches, it manages to describe it partly but fails to fully get all diagrams depicted. While it describes the "red line overlay", it does not identify it as filter on the histogram - even though I tried different prompts trying to guide it towards it.
A simple hand-drawn sketch is described correctly. It can read labels and accurately state the order of the bars depicted.
Occasionally, it does create images even though I did not ask for that. (I mean, hey, it seems to be very proud it can generate images in-chat so no wonder it wants to show off.)