Public Lab is an interesting clearinghouse for DIY “citizen science” tools and information. The general idea is to make available open source instruments and datasets to enable ordinary people to collect, analyze, and disseminate environmental data. I don’t know how “scientific” this might be in all cases, that depends on how you go about it. But it is certainly putting tools in the hands of the people.
“Public Lab is a community where you can learn how to investigate environmental concerns. Using inexpensive DIY techniques, we seek to change how people see the world in environmental, social, and political terms.”
Their offering includes an archive for posting “datasets”, and a bunch of DIY instruments for collecting environmental and biological data.
The Archive
The archive is mostly maps constructed from aerial photography. The archive is curated by Public Labs, who, presumably, require some assurance about the provenance of the imagery.
“To add your open source map to the archive, contact staff@publiclab.org and be prepared to provide some background and to tell the story of your map.”
Not much detail here about this curation process.
The archive is part of the overall political goal of Public Lab , so the datasets may have definitely political purposes. For example, one series of images records coast areas of Louisiana, showing what looks like oil slicks. <<link>> The purpose is explicit:
“This photographic effort and mapping allows us to show that in great detail to a mass audience, and hold BP accountable for its damages,”
This is clearly political propaganda, and the Public Lab tools help make the images look “scientific”, which surely helps the political message.
But this is somewhat dubious science at best. I can’t find any ground truth for these images, to document that the dark spots are, indeed, oil, and from what source (unfortunately, there are a lot of things spilling oil along every coast). Worse, there is no assessment of the actual damage, much of which is not visible in these images. For that matter, without a baseline, how can we assess the damage that might be caused whatever those black splotches are?
Still, these images are dramatic evidence that suggests environmental problems that need to be investigated and handled. Not that we needed more evidence that Louisiana is a mess from the petroleum industry.
The Tools
The DIY tools are much more interesting that the archive. The tools range from the sublime to the ridiculous. There are several devices for assaying water and air quality. These are quite clever, and might be used to monitor your local environment.
On the “ridiculous” end, is a giant selfie stick, used to hoist a camera 5-10 meters overhead, to make a “map” of your garden or similar area. There is nothing wrong with this idea, it’s just kind of, well, not original.
On the more sublime side, the aerial photography is also supported by small blimp and kite systems. Now this is kind of cool, and actually enables some serious reconnaissance (not to mention snooping on neighbors). Add in toy quadcopters, and we’ve got a significant air force, for better or worse.
The various measuring instruments offer cheap ways to measure dust in the air, dirty water, chemicals such as formaldehyde, or other chemicals associated with industrial processes. While much of the documentation explains the construction of the device (usually accompanied by cost comparisons to professional labs), the most important stuff is how to use the instruments.
In particular, it is critical to sample in strategically meaningful patterns, and document the sampling very carefully. The readings mean little if we can’t match them to a time and place, and the meaning can only be constructed from looking at data from the whole area.
Many of the projects give examples of how to organize these efforts, and Public Labs is a digital community dedicated to helping people organize this kind of project. These are critically important features.
Use With Care
I’m impressed with the array of tools, and the overall open source model of Public Lab. I sympathize with their motivation, and much of the work is a good model for any local community to follow.
I remain cautious about over interpretation of environmental data, especially from sparsely deployed, low cost sensors. Most data is ambiguous and noisy, and it is easy to see what you want to find in the data.
This is particularly a challenge when it is your own home and family that may be in danger, and if you go in assuming that “they” are covering up evil doing. Going in with a determination to “show that in great detail to a mass audience, and hold BP accountable for its damages”, you are not well positioned mentally to carefully evaluate the evidence.
It is also important to think globally. These highly localized measurements may not be especially helpful in actually addressing the challenges. For example, documenting water quality issues in one locality is important, but fixing the problem must deal with the whole watershed and especially what is happening upstream.
A Missing Piece: Public Review
Conventional science uses peer review, independent replication, and general devils-advocacy to assure the quality of data and inferences. Perhaps Public Lab might consider something in the way of a review process, with “points” awarded for finding errors and alternative explanations. Anything that survives a thorough criticism will surely have more credibility.
The archive itself should record these reviews and revisions, and there should be a separate section for “five star” datasets that have gone through extensive peer review.
This process can be time consuming and expensive, especially when it drives you to go back and do some more measurements in order to fill in a hole or rule out an alternative explanation. But perhaps using something iike Loomio, it might be possible to create a light weight process that can recruit sharp eyes and minds around the globe to help validate these datasets.
By the way, this is one way that conventional science targets future resources: when a review indicates that current data raises critical unanswered questions, this motivates additional studies. Discussions at Public Labs might generate well specified calls for “most wanted” datasets.
These discussions would also be a channel for cooperation with other parties, including professional scientists who might help connect the dots and shore up the findings with additional data and theory.
A Good Place To Start
Despite my reservations, this is a very useful set of tools, and a good place to start.
Robert McGrath's Blog 