A new study published in Big Earth Data demonstrates that integrating Twitter data with deep learning techniques can significantly improve traffic event detection. By combining semantic, spatial, and temporal features of tweets in a transformer-based multi-feature fusion framework, researchers were able to classify traffic incidents more accurately and identify patterns in real-time. The approach achieved up to 7.93% improvement in accuracy for distinguishing traffic incidents from non-traffic content and 6.09% improvement across three traffic-related categories. These findings highlight the potential of social media as a timely, low-cost source of traffic information for faster response and improved transportation management.

The Alfred Wegener Institute (AWI) and the University of Bremen had already secured extensive data sets from the USA in 2025. The Deutsche Forschungsgemeinschaft (German Research Foundation, DFG) has now approved around 860,000 euros in funding to systematically identify, secure and store endangered data over the long term, based on the PANGAEA data platform.

The tool demonstrates why blowing dust from Great Salt Lake is a significant concern, and why lake levels are a major factor in controlling dust. The tool makes the lake’s changing water depth visually interactive, along with the soil erodibility, affected census tracts, and relevant dust sources that can contribute to unhealthy air quality.

The U.S. Department of Energy’s (DOE) Oak Ridge National Laboratory has released a new version of its advanced slicing software, ORNL Slicer 2, which accelerates and simplifies the creation of large 3D printed parts.