Data-Driven Planet #32 — How Air Pollution Is Predicted
We all share the air we breathe, and this week we take a look at the latest advancements in predicting air pollution. The last 25 years have completely changed the way we predict environmental events, and the next step will be smart systems telling us what to do in order to reduce damage.
“Much of the analysis of the model results — what they meant, how they compared to measured data, etc. — was done manually with human intelligence, reams of printouts, and various custom scripts.”
— Jay Hardikar, IBM researcher, on how data analysis was done 25 years ago.
Ninety-two percent of the world’s population lives in areas with unhealthy air. With the recent news on the devastating effects of air pollution, it’s more important than ever to be able to predict and combat air pollution.
For years, IBM researchers have been working on an intelligent system that is capable of learning to predict the severity of air pollution, with the end goal of providing specific recommendations for how to reduce pollution to an acceptable level.
This is what makes the title story by IBM researcher, Jay Hardikar, such a fascinating read. He explains how he used to work on predicting air pollution 25 years ago and compares it to how it is done in the current age of cloud computing, big data analytics, and the IOT, providing insight into the methods that can be implemented for various types of environmental analysis.
Geospatial Media and Communications has released a free report highlighting the trends and directions of the global geospatial industry along with an index on the geospatial readiness of 50 countries across the globe.
There is also a shorter summary of the report focusing on what cross linkages geospatial has with other technologies.
Thanks to the increasing number of energy sources and growth in energy-related data, energy is no longer considered a line item cost, but a resource that can be optimized and reduced with benefit to the enterprise. But the energy market is fragmented, and different utilities have widely different data formats and release cycles. This requires a lot of data integration effort from the enterprise customers who are looking to benefit from data analytics.
Given the substantial opportunities that advanced energy management provides, the aggregation and accessibility of energy data is a growing business. Therefore, the energy-data-as-a-service model holds great promise for the industry. It enables various industry players to focus on the core offerings and let a third party provide the foundational energy data that enables them to deliver their products and services. Energy optimization and energy procurement are some of the examples of existing services within the energy industry that could benefit from a single source of energy data.
A new library, called GWIS (Graphing Water Information System), can create time-series plots of information measured at U.S. Geological Survey hydrologic data collection sites across the United States. This is a particularly timely announcement given the recent rainstorms in California and elsewhere.
The main GWIS plot features include interactive zoom and legend, custom controls, multiple data series support, and more.
Other Earth data news
Thanks for reading! Are you a fan of weather / climate / environmental data? In Planet OS we have built a open data catalog and APIs for discovering and accessing sensor data. I suggest to give a look, especially if you’re thinking to build your own data-driven application.
Pro tip: My team distributes these blog posts once a week also via email. Simply add yourself to the list and find the next issue in your mailbox.