Air pollution @ EcoHackSF

Transcript

1. Air Quality / Data Quality Collection Validation Presentation Action

2. Air Quality –> Health

3. How is it Quantified? US EPA uses 5 major pollutants

   to calculate Air Quality Index (rough threshold measurements shown) ! • Carbon Monoxide (CO) (9 ppM) • Nitrogen Dioxide (NO2) (52 ppB) • Sulfur Dioxide (SO2) (75 ppB) • Ozone (O3) (0.75 ppM) • Particulate (PM 2.5/10) (12/150 ug/m3)

4. Current Approach to Monitoring • EPA Air Quality System and

   AirNow

5. Resolution Problem • Roughly 4000 active monitoring stations active around

   the USA ! • Average 1 per 1000 miles2 ! • Pollutants such as CO and NO2 can vary widely on the order of meters

6. Problem Illustrated Locally Red Dots are the 3 Official Monitoring

   Stations

7. Improving Resolution • More Sensors? – Cheaper – Easier to

   Install • Moving Sensors? – Track GPS Coordinates – How/Where to go? • Issues of Reliability (Data Quality)

8. General Approach Stepping Back a Bit

9. Who’s Involved • Everyone • Hobbyist/DIY • Grassroots ORGs •

   Universities/Research Institutes • Local Gov’t • Private/Corporations • Federal Gov’t

10. Data Collection Abstracted Sensing Device Local Storage Database Analysis/ Display

    Filtering Use

11. One Concrete Example EPA Server

12. Another Concrete Example Public Mapping System Public Database

13. Sensor Examples • Arduino/DIY • Commercial/Industrial Sensors • Gov’t/Scientific Sensors

    ! • Cheap sensors typically for short-term danger levels • Better sensors can track long-term health risks

14. Mobile Applications • Location Services Integrated (varying quality) • Short-term

    storage • Initial filtering/analysis • Uplink to servers • Real-time feedback • Control sensor settings • Ubiquitous

15. Database Examples • Xively • ManyData • ThingSpeak • ….

    • EPA AirData • BAAQMD • …. • AWS • Your own server

16. Analysis/Presentation • Google Maps • Open Street Map • Various

    GIS systems/protocols • Twitter • Alarm system • Print • ….

17. EcoHack Proposal Focusing In A Little Bit

18. Sensor • Lots of sensors out there – all with

    different performance • Why start with CO? – High concentrations – Can vary widely by location – Immediate & long term toxic effects

19. Mobile Data Collection • Recording Data – Header – Long/Lat

    Coordinates – Time Stamp – Values • Recording Recording Devices – UUID – Calibration Data – Phone serial # and GPS data • Processing on mobile – Filter bad data – Upload in different formats for different servers

20. Thanks West Oakland Environmental Indicators Project