Week 2 Lab Overview

Objectives

In lab this week, we will perform the following:

  • identify a candidate (set of) question(s) to analyze with public Purple Air data;
  • collect (additional) data (as needed) from PurpleAir!

We are working toward the climate/air quality group poster described here.

Generating group datasets with PurpleAir data

For this laboratory project, you will have several decisions to make as a community:

  • What question do we seek to explore with PurpleAir data? What variable(s) are we particularly interested in?
  • What are potential factors that may “confound” our observations?
  • What is an appropriate scale for sampling data across PurpleAir monitoring stations?
  • How many replicates should we take of our data? In this context, replicates could mean the combination of distance to a location/source, time duration, and variable measured if these are variables of interest to your group.

This resource from the Royal Geographical Society may provide additional insight into experimental design elements such as transects. You may be wondering - what is a transect? It is a common method of field sampling. For transect sampling, one typically takes data at specific, equally-spaced locations along a gradient or across (micro)habitats.

This is also a form of systematic sampling, as you would be taking measurements at locations based on a fixed pattern. For more background on sampling procedures and guidance on one way to think through sampling design, please visit this Environmental Chemical Analysis site. Note that not all of the examples there are relevant (e.g. biodiversity sampling, soil sampling), but the general principles in Sections 2.1, 2.2 (our samples from PurpleAir are more akin to a “composite sample”, while the “snapshot” measurements taken with the PocketLab Air units are more akin to a “grab sample”), and 2.4 are pertinent.

A note on replicating data

How might you replicate data collection to try and account for uncertainty in individual estimates? Generally there are two broad ways to achieve this: spatial replication and temporal replication. In your small groups, please discuss the salient variables that could affect your group’s question (e.g. distance to a road or a construction site for pollutants, tree cover for a heat island effect, and time of day).

Additional air pollution information

Note that many consumer-grade air quality units (including PocketLab Air or PurpleAir monitors) report AQI, PM 2.5, PM 10, and O3 data. You can toggle back to the previous week’s Air Quality page to see Myriad Sensors’ (PL Air company) description of how these values are estimated. However, if, for instance, you are interested in the raw values of PM 2.5 that are reported, you may wonder how these values are converted to and from AQI scores. You can navigate to this AirNow calculator to see how your data maps to AQI. If you would like to see the interpretation of AQI values, please check out this AirNow explainer site. # Example group questions

  • How do PM levels vary seasonally?
  • How does air quality vary as a function of distances to roads or the density of roads in an area?
  • How does air quality vary across different local communities which may have varying socioeconomic attributes?

Uploading your data

To ensure that we can help debug any issues that might arise with your data, please have a member of your lab group:

  1. Create a sub-folder for your group in this Week 2 folder
  2. Upload your PurpleAir data to that folder in a CSV format