Internet Could Act As Ecological Early Warning System

Wired is reporting that ecologists think the internet could act as an early ecological warning system based on data mining human interactions. While much of this work has been based on systems like Google Flu Trends, the system will remain largely theoretical for the near future. "The six billion people on Earth are changing the biosphere so quickly that traditional ecological methods can't keep up. Humans, though, are acute observers of their environments and bodies, so scientists are combing through the text and numbers on the Internet in hopes of extracting otherwise unavailable or expensive information. It's more crowd mining than crowd sourcing."

Comments

Predictive Markets

[Anenome]

It's too bad predictive markets have been ruled politically impossible.

If you create a futures market for prediction of future events you give people incentive to share the info they have, and a way to benefit from putting this information on the market, to pay them for sharing knowledge. The Pentagon tried to create such a market for terrorist attacks a number of years ago and the political establishment caught wind of it and murdered it in its crib, calling it a cynical way to make money from terrorist attacks. The truth is, it's a damn effective way to get people to share rare but critical info and no other mechanism is as effective.

This reminds me of an old quote that is apropos in today's political and news-climate: "We should never choose what we want to be true over what we know to be true." Politics shouldn't kill things that produce results.

So, if you want to predict ecological disasters, nothing is stopping us from doing things in a similar way. Some lonely janitor in a nuclear facility with inside info about how poorly the place is being run could bet a thousand dollars in the ecological futures market that said reactor was going to experience an environmental problem within the next year and that would quickly become a sign, as others in the know bet the same way, that the place needs attention and needs it now.

Re:Predictive Markets

[artor3]

So the better method is to shut our eyes really tight, and trust the markets with everything, even when we know just out fallible they are?

We can't stop people from murdering either. Doesn't mean we shouldn't try.

Re:Predictive Markets

[Anenome]

Not exactly. The problem is how to provide incentive for those with rare but useful information to spread that info in society. Doesn't matter what topic we're worried about, that is a problem all of society has. The answer to that problem is a futures market on events.

As for manipulating the markets... lol. Let's say our janitor does bet that reactor 5 is currently leaking. If that's true there's no way to manipulate it. All that can be done is to take his bet. If someone takes his bet, buys his future, that simply generates more interest in that event. Since the janitor knows the reactor is leaking, and someone has taken his bet, he's likely to clap his hands in joy and bet more, which leads to more interest, more activity, etc.

As a regulator you track activity, it doesn't really matter whether the market has an event going up or down.

What actually happens when a futures market on events occurs is that multiple people with independent info about an event occurring can bet it's going to occur. Those who disagree take that bet. When lots of people bet something's gonna happen and no one bets against it, chances are that's good info about an event about to happen.

How exactly do you manipulate this, then? Bet millions of dollars on fake events occurring? I suppose it's possible, if you're willing to lose the money. However, you'd be just one person. Do you then create multiple fake personalities to bet in a similar way? So you draw attention to a non-event, but you can't achieve a groundswell, as happens with a natural event. And you cannot manipulate a market into deflecting attention, you can only draw attention.

I'm not an expert on these types of markets, however, just know that they do work extremely well for what they're designed to do: place value on rare and useful information that would help society if it were known. The market offers no way to subvert that process.

Too quick to keep up?

[gedrin]

Most climate change theories put the temperature shift at fractions of a degree every decade. But this is too fast to keep up with the information? Most areas with extensive internet are already under 24X7 satellite surveillance that can measure temperature and precipitation. What are they expecting to get out of this? If I'm sick, I search for "flu remedy". I will likely be sick tomorrow. It is possible I've spread germs and others will get sick. A new trend of searches for "flu remedy" is a decent indicator that people are concerned about the flu, and that is a decent indicator that they know sick people. If I'm cold when I go out, I search for "winter coat". It may or may not be cold tomorrow, but probably will be since cold days group together. It is likely other people are also cold. But this isn't a predictive indicator. I'm cold because the temp already has dropped. The weatherman knows this, and I may have done my search because he said it was going to be cold tomorrow. The satellite knows this. Vast records are kept about how cold it was in Podunk, Kansas. You could predict that coat searches will go up because it was cold, but once people are searching for coats it's as a result of data you've already got from much more scientific measurements. Even if you could find a method that was predictive, what indication is there that populations, even if good at predicting short term weather (I'm not convinced), are competent in any way at predicting long term climate changes? Are they really arguing that the people, many of whom have a disturbing habit of living on flood plains, barrier islands and below sea level, are going to produce accurate data about future environmental shifts? It seems far more likely that this is a chance for someone to do research using a cool new toy. Our climate models current are pretty universal in their failure to predict the future, and that's with tons of solid data. This seems like a buzzword filled hole designed to sound urgent and cutting edge. An unproven model using an unproven data collection method can be used to justify any conclusion that will keep this project funded as long as it sounds cool enough to the people footing the bill.