Move over millibars, Flickr is the new atmospheric metric for hurricane central pressure

From the University of Warwick, and the ultra-short baseline department comes this study of ridiculous proportions. I don’t have the words to accurately describe this utter waste of time and money. I can just see future NOAA bulletins:

HURRICANE WHOPPER FORECAST/ADVISORY NUMBER  18

NWS TPC/NATIONAL HURRICANE CENTER MIAMI FL   AL122005

2100Z SAT AUG 27 2015

AT 4 PM CDT…2100Z…THE HURRICANE WATCH IS EXTENDED WESTWARD TO

INTRACOASTAL CITY LOUISIANA AND EASTWARD TO THE FLORIDA-ALABAMA BORDER.  A HURRICANE WATCH IS NOW IN EFFECT ALONG THE NORTHERN GULF COAST FROM INTRACOASTAL CITY TO THE ALABAMA-FLORIDA BORDER.

GIVEN THE ACTIVITY MEASURED ON FACEBOOK, TWITTER, AND FLICKR A HURRICANE WARNING WILL LIKELY BE REQUIRED FOR PORTIONS OF THE

NORTHERN GULF COAST LATER TONIGHT OR SUNDAY.  INTERESTS IN THIS AREA SHOULD MONITOR THE PROGRESS OF WHOPPER AND POST PICTURES OF ANY OBSERVED STORM ACTIVITY TO HELP NHC GAUGE CENTRAL PRESSURE.

Hurricane Sandy’s impact measured by millions of Flickr pictures

A new study has discovered a striking connection between the number of pictures of Hurricane Sandy posted on Flickr and the atmospheric pressure in New Jersey as the hurricane crashed through the US state in 2012.

Hurricane Sandy was the second-costliest hurricane to hit the US, hitting 24 states in late October last year, with New Jersey one of the worst affected.

In 2012 32 million photos were posted on image hosting website Flickr and by counting the number of pictures tagged either ‘Hurricane Sandy’, ‘hurricane’ or ‘sandy’ between October 20 and November 20 2012, a team of researchers led by two Warwick Business School academics, Tobias Preis, Associate Professor of Behavioural Science and Finance, and Suzy Moat, Assistant Professor of Behavioural Science, found a strong link to atmospheric pressure dropping in New Jersey.

In fact, the highest number of pictures posted were taken in the same hour in which Hurricane Sandy made landfall in New Jersey. In Quantifying the Digital Traces of Hurricane Sandy on Flickr, to be published in Scientific Reports today (Tuesday November 5), Tobias Preis and Suzy Moat, of Warwick Business School, Steven Bishop and Philip Treleaven, of UCL, and H. Eugene Stanley, of Boston University, suggest that using such online indicators could help governments measure the impact of disasters.

Preis and Moat’s work has previously uncovered a range of intriguing links between what people look for online and their behaviour in the real world. Recent results revealed that changes in how frequently people searched for financial information on Google and Wikipedia could be interpreted as early signs of stock market moves, and that internet users in countries with a higher per capita GDP search for more information about the future.

“Our steadily increasing use of digital technology is opening up new and fruitful ways to document and follow human actions,” said Dr Preis. “Building on our recent work, we asked whether data from photos uploaded to Flickr could have been used to measure the impact of Hurricane Sandy.

“Our new results show that the greatest number of photos taken with Flickr titles, descriptions or tags including the words ‘hurricane’, ‘sandy’ or ‘Hurricane Sandy’ were taken in exactly the hour which Hurricane Sandy made landfall in New Jersey.

“Examination of the number of Hurricane Sandy related photos taken before and after landfall reveals a striking correlation with environmental measurements of the development of the hurricane.” Dr Moat added: “As the severity of a hurricane in a given area increases, atmospheric pressure drops. We found that as atmospheric pressure in New Jersey fell the number of photos taken rose and as atmospheric pressure climbed again the number of photos taken fell. (See graphs attached)

“Plotting the data revealed that the number of photos taken increased continuously while ‘Sandy’ was moving towards the coast of the US. This study would suggest that in cases where no external sensors are available, it may be possible to use the number of Flickr photos relating to a topic to gauge the current level of this category of problems.

“Flickr can be considered as a system of large scale real-time sensors, documenting collective human attention. Increases in Flickr photo counts with particular labels may reveal notable increases in attention to a particular issue, which in some cases may merit further investigation for policy makers.

“Appropriate leverage of such online indicators of large disasters could be useful to policy makers and others charged with emergency crisis management: in particular if no secondary environmental measures are available.”

###

Quantifying the Digital Traces of Hurricane Sandy on Flickr is published in Nature Publishing Group’s Scientific Reports today, at http://www.nature.com/srep/2013/131105/srep03141/full/srep03141.html

This URL will become live when the embargo lifts at 2pm UK time on November 5 2013. Copies of the paper ahead of the embargo can also be requested from Tobias.Preis@wbs.ac.uk.

Advertisements

  Subscribe  
newest oldest most voted
Notify of
John V. Wright

That’s it. Words have finally failed me. #ifthesearebusinessschoolacademicsweareallupshitcreek

Oldseadog

I’ve been asleep for months and now it is the first of April.

elftone

Wow, what a load of useless b*llocks. These people have no clue about cause and effect. Next, they’ll be predicting extreme weather events using the Kardashian metric. Idiots.

Sweet Old Bob

Positive feedback? Big storm. Big media hype.Big response.Maybe all these pictures caused Sandy./sarc.

JimS

If they found that online searches increased by the millions for SUV’s, I assume this would predict that people will be buying more SUV’s, and thus, global warming will increase to catastrophic levels. Ok, I got it now.

Wow is right, elftone. What a waste of taxpayer dollars.

Eliza

National Centre for Environmetal prediction forecasts are amazingly accurate up to 7 days in advance. One good bit of news for the modelers (which I am guessing are mostly meteorologists). I have been observing these maps closely for about a year now.
http://wxmaps.org/pix/sa.850.html
Basically you can be about 100% sure if it will rain next sunday or not. Congrats to the NCEP!

Alan Robertson

Ok, this is two absurdities in one day. We’ve seen it before. What’s the record, six/day?

steven

Did I just accidentally stumble upon “The Onion” instead of WUWT?

RCF

So, in cases where no external sensors are available, we’ll rely on individuals having internet connections in order to gauge the severity of the event. Sounds like a plan…

Gary Hladik

I think Michael Mann may have a new proxy…

Gene Selkov

Flickr is a site for photographers — people who work on their pictures before posting them. I’m still posting my 2008′ photos, with several thousands in the pipeline.
I may be an outlier, but the typical delay between taking a picture and posting it on Flickr is several days. So much for advance warnings.

Post normal science
Post on Flickr
Post hoc ergo propter hoc
The climate is gonna go postal!

Robert Clemenzi

You may consider this to be an “utter waste of time and money”, but there are a number of groups in the Washington DC area, including DARPA and various news agencies, actively investigating how to use social media on the web to obtain information that is otherwise difficult to obtain.

So they aren’t doing no proper education in universities if they can’t even teach the basics of science such as correlation does not mean causation. init.

Resourceguy

Or more likely, the short attention span, smart phone society has imposed itself into research efforts and popular science.

BradProp1

The warmists obviously have no need for Ex-lax being they can pull stuff like this out of their asses!

DirkH

Robert Clemenzi says:
November 5, 2013 at 8:46 am
“You may consider this to be an “utter waste of time and money”, but there are a number of groups in the Washington DC area, including DARPA and various news agencies, actively investigating how to use social media on the web to obtain information that is otherwise difficult to obtain.”
So you’re implying that atmospheric pressure is difficult to obtain without social media. I disagree. In my opinion, a barometer excels at obtaining atmospheric pressure, and beats the NSA in this regard. A big problem in obtaining correct atmospheric pressure readings from social media that still awaits a solution is for instance sarcasm; a topic that the barometer elegantly sidesteps by direct measurement of atmospheric pressure.

davidmhoffer

Whoa up folks.
They aren’t claiming that their methodology predicts the hurricane. They’re claiming that their methodology may be of value in terms of understanding extent and extreme of an event in real-time. While I think this study is rather narrow in that it looks only at Flickr, leveraging the enormous amount of data the arrives in real-time from social media such as Flickr, Twitter, Facebook and others has merit.
Not only does this technique work, it is already in use for other purposes. When a major consumer goods manufacturer makes an announcement of, for an example, a price increase, they can scrape an enormous amount of information from social media in a matter of hours to learn how much or little that price increase is affecting them in the market. When a new product (say a new whiz bang cell phone) has an unexpected problem, the manufacturer will know within hours, perhaps minutes, how extensive the problem is, if it is constrained to a specific geography, it it appears only in combination with something else (ie only a specific telco) and many other things.
Leveraging social media to understand in real-time which areas are most heavily hit, and by what (fallen trees, flooding, fire, what?) could be of immense value in emergency response.
We have a propensity to dismiss out of hand studies like this as useful predictors of anything, and correctly so. But applied to what is happening and where it is happening, and what should be done about it has merit.

New study shows victoria secret models wear less than at any time in
History. Must be getting warmer…..sorry

Old'un

Bob Tinsdale says at 8.43
‘what awaste of taxpayers dollars’ .
Unfortunately for we Brits Bob, they are probably taxpayers Pounds on this occasion!

LeeHarvey

@ Gary Hladik –
Bravo, Sir.

Jim Clarke

We have been arguing for years about the surface temperatures record because we can not quantify the small ways in which any given temperature recorder might be inaccurate or not a true representation of what has actually occurred. So how can a completely unregulated, variable laden metric like Flickr photos be anything but a blank canvas on which to draw any conclusion one desires?

I need to know the weather tomorro ,should I turn on the weather channel or facebook?

Tomorrow …..

Joe

Tobias Preis and Suzy Moat, of Warwick Business School, Steven Bishop and Philip Treleaven, of UCL, and H. Eugene Stanley, of Boston University, suggest that using such online indicators could help governments measure the impact of disasters.
————————————————————————————————————-
Don’t be too dismissive.
Only last week I noticed on Facebook that lots of my friends (over 10%) were changing their profile pictures to photos of giraffes. Sure enough, when I investigated a little further, it turned out that they had all turned into giraffes for 3 days having got a riddle wrong.
Would this disaster have been noticed by anyone without the photo changes?

Eric in Escondido

Unfortunately, this is probably not the silliest study to have been published by such “academics.”
“Meteor predicted to hit NY City!” Flikr traffic zooms. Correlation!

TerryS

Re: davidmhoffer

They aren’t claiming that their methodology predicts the hurricane. They’re claiming that their methodology may be of value in terms of understanding extent and extreme of an event in real-time.

I think it more likely that what they measured was the extent and extreme of the media hyperbole (in real time) regarding hurricane Sandy.

Owen in GA

davidmhoffer says:
November 5, 2013 at 8:59 am
Whoa up folks.

Or they could be finding that when a storm hits an area with a very high population density that more people post photos of said event on social media. I think there should be a common sense requirement on these papers – are they not peer reviewed?

Steve Oregon

What we are witnessing is beneath stupid or any other mental deficiency, immaturity or affliction.
It is CCGSI
Catastrophic Collapse of Global Scientific Integrity

Steve from Rockwood

The trick is to develop an effective Kim Kardashian filter.

Gee, I wonder what the correlation with Hurricane Andrew was with the posting of FlickR pictures?
I am sorry, while I am no meteorologist, it staggers my mind that someone would actually try to link pictures to intensity. Especially since Social Media on the internet is so new coupled with the fact that cameras with cell phones were not the “norm” even when it did start up a few years ago!
Guess we are screwed. I suspect every hurricane is going to get lots of postings in the future. Which should increase the intensity no doubt.

heaterguy

It’s amazing!! They’ve developed a new way to stop all future hurricane activity. Disconnect Flickr!

Louis Hooffstetter

…by counting the number of pictures tagged.. ‘hurricane’ or ‘sandy’… a team of researchers…found a strong link to atmospheric pressure dropping in New Jersey.
…by counting the number of pictures tagged.. ‘Kardashian’ or ‘Kanye’… a team of researchers…found a strong link to IQs dropping in New Jersey.
Wow, this research is easy!

R. de Haan

Flickr, just another measure of mass hysteria.

Louis Hooffstetter

…by counting the number of video recordings showing President Obama.. ‘promising 300 million Americans that “if you like your health-care plan, you will be able to keep your health-care plan, period.”… a team of researchers…found a strong link to IQs dropping in America.
Sorry, I just couldn’t resist.

jorgekafkazar

davidmhoffer is perfectly correct. i don’t think most readers here grasp the intent of the method. It’s broader than weather. Any event covering a large area or social unit can be indirectly assessed with this sort of proxy, giving interested parties an idea of the event’s seriousness without having to first install instruments of any sort. Do you seriously doubt that there are people whose jobs require that they hourly google, for just one example, “împëächmëñt?” Flickr provides a way to count crowds, pitchforks, torches, etc., as well. Very useful.

Aphan

They highlight a relationship between increasing and decreasing “atmospheric pressure” to the increase and decrease in the number of photos being posted why? To classify the study as being related to climate science? Because air pressure controls human urges to post photos? To get it published in Nature rather than a social science magazine? Did they mention global warming or climate change in their abstract? If so, Nutt and Cook can add it to their consensus results….

David in Cal

This “study” is ridiculous, but something like it could make sense. I attended a recent talk at Google. The Google-employed researcher demonstrated a link between the number of google searches (using various input terms) and weekly unemployment figures. These would allow a person to estimate the next weekly figure a few days before it was reported.

So what is the “Behavioral Psychology” behind college professors who obsess over minutiae to the degree that they see significance in trying to relate counts of pictures to weather phenomena?

D.J. Hawkins

@davidmhoffer says:
November 5, 2013 at 8:59 am
Thanks, I thought for a moment I was reading responses to some post other than the one at the top of my browser. Knee jerk reactions and failure in reading comprehension ain’t pretty whether it’s on the CAGW side or the skeptic side, but I thought we were at least a little more contemplative.

Txomin

Yes, it’s ridiculous beyond belief. But let’s not forget that the authors are likely to be proud of their work, ok? Enough stomping on people’s dreams.

Tom O

davidmhoffer says:
November 5, 2013 at 8:59 am
“Whoa up folks.
……
Leveraging social media to understand in real-time which areas are most heavily hit, and by what (fallen trees, flooding, fire, what?) could be of immense value in emergency response. ”
Yes, everywhere you go there are thousands of smart phone users out there taking pictures of things happening in the middle of the storm so someone doing research on the right set of words in real time could determine where places are hardest hit. Excuse me, but wouldn’t it have been more productive to take that smart phone and, say, dial 9-1-1 when you have a real problem in your area, rather than tweeting or posting pictures on social media? I think so, are you sure you don’t?

See - owe to Rich

And of course all those people Flickring about Hurricane Sandy didn’t even realize…there were no sustained hurricane force winds over land in the eastern USA at any time! I mean, at any time during Sandy’s existence.
And so we don’t need NOAA or NWS measurements any more to determine if a storm is a hurricane, we just crowd-source the answer. This way the drought in major hurricanes landfalling the USA can be broken!
Rich.

davidmhoffer

Owen in GA;
Or they could be finding that when a storm hits an area with a very high population density that more people post photos of said event on social media.
>>>>>>>>>>>>>>>>
You of course have to adjust for population density. But if in the middle of a disaster you get 1,000 twitters with the word “fire” in them all from within a two block radius, there’s a pretty darn good chance that there is a major fire in that area. If photo uploads on flickr and facebook show what is burning, say a 50 story building versus 50 houses versus a manufacturing plant that makes fireworks, would that information not be of value to emergency dispatchers and first responders on their way?
jorgekafkazar says:
November 5, 2013 at 10:05 am
davidmhoffer is perfectly correct. i don’t think most readers here grasp the intent of the method. It’s broader than weather. Any event covering a large area or social unit can be indirectly assessed with this sort of proxy
>>>>>>>>>>>>>>>>>>>
Thanks for your support. In fact, this is already being used in business and media. Pay attention to news reports of major events for example. Increasingly, we see real-time reporting of what is going on at the scene via data scraped from social media. All one has to do to see that this is an emerging and effective technique with real possibilities for emergency response is to follow the news of the next major event and pay attention to the sources of information and the speed with which they get reported.

Mike Bromley the Kurd

“Building on our recent work”…….
Now there’s a stretch.

Jimbo

Was this published in the Journal of Improbable Research?

Aphan

So hilarious I read it again…”as the storm becomes more severe….atmospheric pressure drops…and the number of photos increased at the exact same time”. Wow. The amount of DAMAGE increasing, thus making for more dramatic photos that one would want to capture and share….wasn’t the obvious choice? And as the storm moves on….and damage/drama decreases, thus lowering interest in taking photos….nope. All caused by atmospheric pressure. Idiots.
Jorgek and davidh-
Funny thing is….people connected to social media, by default are connected to the internet…and thus all digital NEWS media. In the week prior to Hurricane Sandy, I noted that news about the approaching storm increased steadily irregardless of the storms atmosperic pressure. And I live nowhere near the East Coast. From that fact alone, I can postulate that his study is about an event that had MULTIPLE, OUTSIDE, NATIONAL “sensors” and in no way indicates that Flickr itself was a “sensor” independent of any other large scale sensors. From this fact, i can extrapolate that the more hype the storm got, the more people began watching for its effects. I postulate that the worse it got, the more people reached for their cameras because they found the chaos to be out of the ordinary, and worth capturing etc. This study tries to paint the Flickr activity as a precursive indicator that reflected the photographers response to “atmosoheric pressure changes”. Come on!
Since the people in US East Coast cities are highly digital, and highly internet savvy, of COURSE they are going to document and share digital data. But large scale events have ALWAYS been documented with photographic evidence….”without having to install censors first”.

Crispin in Waterloo

Mike Bromley the Kurd says:
>>“Building on our recent work”…….
>Now there’s a stretch.
Exactly. Building what? With me, it would be their reputation, of sorts.

Would Hurricane Whopper count as male or female? Or are we now entering the era of gender-neutral hurricanes?