This week’s list of top data news highlights covers April 5, 2025 to April 11, 2025 and includes articles on piloting humanoids for household tasks and using machine learning to save endangered marine species.
The city of Philadelphia has launched an open data initiative called Philly Stat 360 to improve public access to information on citywide road safety efforts. The platform includes 14 datasets with interactive maps that show where the city has installed speed bumps, the locations of serious or fatal crashes over the past year, and updates on ongoing projects along high-traffic corridors. The data reveals a 71 percent increase in average daily bike trips since 2019 and a 21 percent decrease in fatal and serious crashes on Roosevelt Boulevard, one of Philadelphia’s most dangerous roads.
UK Power Networks, an electricity distributor, is testing smart heat pumps that use AI to run more efficiently and support the electricity grid. These systems pull heat from the air or ground to warm homes, and the AI system helps decide when to run them based on real-time demand, temperature, and usage patterns. The goal is to reduce strain on the grid, avoid costly infrastructure upgrades, and keep homes warm without increasing energy bills. In early trials, they’ve successfully shifted energy use away from peak hours, cutting demand during the busiest grid periods by up to 40 percent without affecting comfort.
3. Enabling Autonomous Food Delivery
Uber Eats is partnering with Serve Robotics, an autonomous sidewalk delivery company, to deliver food autonomously in the Dallas-Fort Worth metro area. The robots use a range of sensors along with AI tools to identify objects and avoid obstacles in real-time. Serve’s robots have completed thousands of emissions-free deliveries with zero emissions, offering a lower-impact alternative to car-based delivery in dense urban areas.
Verge Genomics, a California-based biotech company, has used AI sensor data from 54 patients with amyotrophic lateral sclerosis (ALS), a neurodegenerative disease. Over an eight-week period, the system detected measurable changes in walking speed, how often patients turned in bed, variability in breathing, sleep quality, and the time it took to complete speech tasks. The results show that AI tools can track the progression of ALS across multiple functions without relying on in-person clinic visits.
FitXR, a virtual fitness company based in the UK, has developed a VR app that uses motion tracking to monitor users’ movements, pace, and consistency during workouts. An integrated AI system analyzes this data in real time to adjust difficulty levels and recommend new activities, including Muay Thai and Brazilian Jiu-Jitsu. FitXR reports that users complete longer sessions and return more frequently compared to earlier versions of the app.
6. Predicting Quantum System Dynamics
Researchers from Hong Kong University of Science and Technology, Shenzhen University, and the Quantum Science Center in China have developed an AI model that learns to predict how quantum systems evolve by analyzing sequences of measured outputs, such as spin orientation and gate performance. Using a type of neural network trained on this data, the model can inform strategies to help quantum computers stay stable during operation, making it easier to correct errors, calibrate hardware, and improve overall reliability. In tests, it achieved over 99 percent accuracy.
1X, an AI and robotics company based in Norway, has developed a humanoid robot named “Neo” that performs household tasks. Neo is remotely operated by a human operator through a VR headset and controllers, allowing it to perform actions like folding laundry or loading a dishwasher. While the robot can walk and avoid obstacles on its own, it relies entirely on human input to complete tasks. 1X is using data from these sessions to train Neo’s AI system, with the long-term goal of making the robot more autonomous over time.
8. Monitoring Endangered Species
Researchers at the University of the Highlands and Islands in Scotland have partnered, in collaboration with government agency NatureScot, have launched Rate a Skate, an AI-powered tool for identifying and monitoring the critically endangered flapper skate in UK waters. The system uses a neural network trained on thousands of images of the marine animal to recognize individual skates by the unique spot patterns on their backs. Underwater cameras deployed capture footage that helps researchers track population numbers and study behavior, providing data to guide conservation efforts.
A Latin American port services operator is using a VR training system in Chile to track how well crane operators perform in simulations. The platform replicates port environments, including weather variations, ship movements, and heavy machinery operations. The system collects data on reaction times, error rates, and task efficiency, which are then analyzed to tailor training programs and improve operational readiness based on real-world scenarios.
10. Mapping Groundwater Reserves
Researchers at the University of Arizona have developed HydroGEN, an open data tool that monitors and forecasts groundwater flow nationwide. HydroGEN uses a computer simulation of the water cycle to model how water moves both underground and on the surface. An AI algorithm then adjusts this simulation based on historical data, ultimately predicting where and when groundwater will flow. This combined approach provides accurate, timely information that helps farmers and water managers plan for water availability.
