Fighting asthma with data

Fighting asthma with data

Government Computer News just published an article on our work with the GASP project entitled Fighting asthma with data.

Year after year, Chattanooga, Tenn., ranks as one of the most challenging cities to live in for people who suffer from asthma and seasonal allergies, according to the Asthma and Allergy Foundation of America. It was eighth on the list for 2015, sixth the year before.

To help the city better target its air quality efforts and give asthma sufferers more information on which areas they should avoid, a team of researchers from the University of Texas at Dallas are working to install highly sensitive sensors to map in real time the areas with the most particulate matter in the air.

The Geolocated Allergen Sensing Platform will help “determine whether real-time allergen and pollution collection and analysis on very fine geographic scales — the scale of a city block or less — can improve health and wellness,” David Lary, the project’s principal investigator, told GCN. “The amount of pollution and allergens we encounter can be orders of magnitude different depending on which route you take.”

GASP aims to create a network of sensors throughout the city to make it possible to know where the most polluted areas are. If Google Maps can give travelers the quickest route to their destination, then GASP may be able to give residents the least polluted route, Lary said.

Currently, particulate matter is measured by sensors provided by the Environmental Protection Agency, he said. Those sensors, while well calibrated, are very expensive, making them impractical for the street-level analysis that would be required for a system like GASP.

The GASP sensor uses a laser to detect and measure the particulates and pollen. Based on research Lary has done in Texas, he estimates that sensors would be needed every half a kilometer to five kilometers.

The brains of the sensor is the Waggle platform, created by the Department of Energy’s Argonne National Laboratory and used in Chicago’s Array of Things. Waggle is an onboard computer that analyzes the data and is capable of machine learning that helps in the calibration of the device.

Waggle will be connected to the city’s high-speed fiber network to provide real-time updates through an application programming interface. A longer-term goal is setting up a dashboard with visualizations, he said.

Right now there is only one sensor deployed in Chattanooga. It was installed last fall to so it could be calibrated during the pollen season and will go through another round of calibration in the spring. Funding from the National Science Foundation will cover eight sensors in all. Lary doesn’t have a hard timeline for when the remaining ones will go up, but said the team is working quickly.

“The combined use of all these technologies — big data, remote sensing, network connectivity, machine learning, the so-called Internet of Things — it’s all very up-and-coming,” Lary said. “It’s an approach that has tremendous potential to have a massive societal impact.”

Breath Analysis

Ivan R. Medvedev, Robert Schueler, Jessica Thomas, Kenneth O, Hyun-Joo Nam, Navneet Sharma, Qian Zhong, David Lary, Philip Raskin (2016), Analysis of Exhaled Human Breath via Terahertz Molecular Spectroscopy, IRMMW-THZ 2016, 41st International Conference on Infrared, Millimeter and Terahertz Waves, 25-30 September 2016, Bella Center, Copenhagen, Denmark.

Abstract—We report on our progress in utilizing THz breath sensing in several bio-medical diagnostic applications. Our work bears promise in applying this technology to non-invasive analysis of blood glucose based on chemical composition of breath, as well as assessment of asthma related airway inflammation. Our most recent testing of CMOS based THz breath sensor, in the evolution of this technology towards compact and affordable implementations, is discussed.

Random Forest Example

Here is an exampleRF using a Random Forest (TreeBagger) in matlab.

The example:

  1. Loads a matlab test dataset.
  2. Finds the capabilities of computer so we can best utilize them.
  3. Trains a TreeBagger (Random Forest).
  4. Creates a scatter diagram.
  5. Estimates the relative importance of the inputs.
  6. Examines how many trees are needed.

A short course on Big Data & Machine Learning

https://bigdata.eventos.cimat.mx

BIG DATA IN SERVICE OF SOCIETY

Machine learning and multiple massive Big Data sets can be of great use for a wide variety of scientific, societal and business applications. The World Health Organization issued a report stating that seven million people died in 2012 from pollution related issues. Each year there are an estimated 219 million cases of Malaria. Eleven states have recently made drought declarations. Every year the US spends between $1 and $2 billion fighting fires. Issues such as these are of massive societal and personal relevance. Big Data and machine learning can provide invaluable tools for both improved understanding and making data driven decisions and policy.

This workshop will give an introduction to a wide range of Big Data applications of major scientific and societal importance such as environmental health, drought and water issues, fire.

The practical tools introduced can be readily used in a wide range of applications from research to real time decision support. The data used comes from a wide variety of sources including scientific instrumentation, social media, remote sensing, aerial vehicles and the internet of things.