Showing posts with label Data Visualization. Show all posts
Showing posts with label Data Visualization. Show all posts

Sunday, May 17, 2009

Sunday, Flu-ey Sunday

(Visualization: Evan Robinson, Group News Blog; Data: WHO | Influenza A(H1N1))

Sunday morning in Paris, and Rue Cler is quiet, especially compared to Friday & Saturday nights.

I'm sorry to have been so long between updates, but the news is relatively benign and we're coming back to the hotel exhausted every day (sometimes twice a day if we're worked really hard). We have covered many of the basics: ile d'citi; tour eiffel, flea markets; seine cruise, etc. Today is for the Louvre and possibly Musee d'Orsay.

But first, I'll try and struggle through our unreliable network and get an update (probably the last) on our current flu outbreak.
(Visualization: Evan Robinson, Group News Blog; Data: WHO | Influenza A(H1N1))

The world has taken this outbreak very seriously. WHO has dusted off their brand-spanking-new pandemic phase system and news, especially television news, has given H1N1 the sort of treatment previously limited to wars and major terrorist attacks.

The current data doesn't justify that level of response, but that doesn't mean that we are completely out of the woods yet.

This outbreak appears to be starting to die out. We've been well under exponential growth rates for a week or so and even the current Ground Zero (the US) has been under exponential growth rates for 4 or 5 days. Unless there is another burst of activity, this phase will end by the end of May or so (possibly the middle of June) and we'll all stick our heads back into the ground for a while.

If, however, this flu follows the pattern of the 1918-1920 pandemic, we can expect a resurgence in the fall, which may well be more lethal (the 1918 flu had varying lethality over different phases, but overall was probably the most lethal influenza ever seen). Several processes can contribute to growth in flu's lethality, including progression (as the virus moves from host to host, sometimes it gains potency), transposant changes (the exchange of DNA material from multiple viral strains in one host), and good old fashioned "mutation". The Great Influenza points out that the 1918 flu, by starting out incredibly lethal, almost had to "revert to the mean" and become more benign, but even the more benign strains of the later phases were still twice as lethal as a "typical" flu season.

Two factors combined in the 1918 flu to make the virus more lethal to young adults (20-40) than the typical flu: a previous pandemic (not the 1889, but another, possibly so weak that it had gone unnoticed) had apparently provided older people some measure of immunity to the 1918 flu; and many victims of the 1918 flu were actually killed (or severely weakened) by an immune response rather than the virus itself -- because young adults have stronger immune systems, they were hit harder than children and older patients.

Ultimately, in 1918 the most effective public health response was isolation. A number of locations or organizations enforced strict quarantine and escaped the flu altogether, but army camps, which were notoriously poorly quarantined, proved to be major elements in spreading the disease. Without antiviral drugs or a real understanding of the infectious agent (scientists argued for years over whether bacillus influenzae was the etiological agent of the 1918 flu pandemic, until a "filterable virus" was eventually determined to be the cause), patients were basically on their own against the disease, and the only useful response was a public health one. Because of Wilson's obsession with the war in France, public health took a backseat to military preparedness and morale boosting, and the appropriate public health measures were not taken quickly or strongly.

Mexico's decision to shut down nearly all public events for a week was exactly the right response to create a firewall against the epidemic: flu spreads most easily in crowded conditions, and with a 24-48 hour incubation period, a week was sufficient to slow the spread of the disease. The US response (which was basically to do nothing) has shown to be inferior as the number of cases in the US has risen. Fortunately, we aren't dealing with an extremely lethal strain. This time.

There's not much else to say unless/until the next wave comes around. It wouldn't be surprising to see another outbreak over the summer or into the fall. Whether this particular flu becomes more lethal or less is an open question, and we'll just have to wait and see.
There's more...

Sunday, May 10, 2009

Tell me on a Sunday

(Visualization: Evan Robinson, Group News Blog; Data: WHO | Influenza A(H1N1))

So here we are in London. I've missed several WHO updates, and the difference between today and Thursday is at least somewhat striking.

Worldwide, we continue to track below exponential levels, but the country breakdown has changed. Mexico is no longer Ground Zero for concern.

The USA is.

US cases roughly doubled in the last reported 12 hours, over Friday night to Saturday morning, jumping from almost 900 to over 1600, and crossing the exponential track from below. Apparently WHO isn't reporting over Saturday afternoon through Sunday, so we'll have to wait for their next update (Monday morning?) to see if this is a one-time leap or the beginning of a very scary trend.

Since deaths aren't tracking up any faster, I continue to believe that we are in pretty good shape for this phase even if the US jumps the exponential track on the high side, but it's something to be concerned about.

In addition, WHO is tracking a couple of active cases of A(H5N1) aka avian flu in humans (one in Egypt confirmed 6 May, one fatal in Vietnam on 22 April). Single cases of avian flu are not necessarily of epidemic concern, but the possibility of a transposant combination of A(H1N1) which spreads easily from human to human and A(H5N1) which has a 30-50% mortality rate is the basic human public health nightmare.
There's more...

Thursday, May 7, 2009

I heard the flu's today, oh boy

(Visualization: Evan Robinson, Group News Blog; Data: WHO | Influenza A(H1N1))

Good news this morning. We've failed to hit an exponential growth curve now for 24 hours (yesterday's morning report through today's morning report). The graph clearly shows Mexican cases and worldwide cases falling off to the right (and below) the exponential growth curve, although US cases are still tracking it very closely.

I remain cautiously optimistic -- and less cautiously about this phase -- about being out of the woods. There is no question that we must remain vigilant and concerned about changes to the virus via mutation, recombination, or progression, but those are issues for lab scientists, epidemiologists, and public health authorities, not the general public. It looks increasingly like we've dodged the pandemic bullet for this piece of the battle.

This is necessarily short because I'm packing for my starter trip to Europe: two weeks in England and France. I'll blog as I'm able, but I don't know how much time, energy, or network I'll have available.
There's more...

Monday, May 4, 2009

More Daily Flu-ness

(Visualization: Evan Robinson, Group News Blog; Data: WHO | Influenza A(H1N1))

Current WHO pandemic phase: 5
Laboratory confirmed cases worldwide: 1085 cases in 22 countries
Laboratory confirmed deaths worldwide: 26 in two countries
Laboratory confirmed cases in US: 286
Laboratory confirmed deaths in US: 1

As rapid testing comes online all over the world we're now seeing exponential growth in confirmed cases. If the exponential growth continues we'll see about 1500 total confirmed cases tomorrow and about 2500 2400 the next day. Since we don't have any real idea of how many cases might be out there but not confirmed, we still have no idea of the actual mortality rate. If we assume that the confirmation rate is identical for cases and for deaths (a dangerous assumption, but I think a relatively conservative one) then the mortality rate is about 2.4%.

Swine flu fears subside, but second wave looms correctly points out that earlier major pandemics involved several waves of morbidity with increased mortality in the fall and winter waves. If we dodge this particular bullet throughout May and June, we'll still have to worry about October and January.

Mexico appears to feel that the epidemic is winding down, while Canada has our first "severe" case. Janet Napolitano has pointed out that currently A(H1N1) "is not stronger than regular seasonal flu".

At this point I'm not too worried. As I've pointed out before, a typical flu season in the US averages about 36,000 deaths, and we've got 26 so far (.0007 times as many). We can, of course, see a rapid increase in cases and in deaths, but we may not, or at least not for a while. It's not time to panic yet. We are seeing tremendous acceleration in confirmation of cases, but we're not seeing a concomitant acceleration in confirmation of deaths. I believe that means we're seeing increased detection, not increased morbidity (sickness).

I've been reading The Great Influenza, and it's fascinating. I recommend it. It begins with a summary of the American medical education system in the 1800s, then provides an overview of the special conditions which contributed to the spread of influenza in the American army camps (the first known cases of 1918 "Spanish" flu were detected in Kansas). I'm looking forward to the rest :-).

[Updated 20090505 1245PDT: changed next day estimate from 2500 to 2400 after taking a more careful look at my graph -- the 2500 was just tossed off and I frankly didn't think the 1500 was going to be accurate enough to care about, but it was so I figured I'd better fix the mistake -- ER]
There's more...

Monday, April 27, 2009

NYTimes Flu Graphic

(Graphic: New York Times, Understanding Swine Flu)

Alongside the Veratect Twitter and Biosurveillance Blog, this looks like a useful tool for keeping up on what's happening with the Mexican flu.
There's more...

Saturday, March 14, 2009

Choosing an Energy Source


I was reading The Transition Handbook in preparation for a workshop next weekend, and there's a couple of great charts showing the carbon intensity (grams of CO2 equivalent produced per kilowatt-hour) and EROEI (Energy Returned on Energy Invested) of various energy sources (although oil is prominent by its absence, I added some in later). Those charts are online at The Ecologist magazine. Although I found them interesting and useful, I thought they would be much more useful if they were combined.

The result is above: a scatter chart showing the relationship between EROEI and Carbon Intensity for various forms of energy. Because a high EROEI is good, we would prefer to use energy sources far to the right. Because low Carbon Intensity is good, we would prefer to use energy sources far down on the chart.

Several energy sources were "outliers" -- they didn't fit nicely on the chart as scaled, so they are marked by half cut-off blue dots at the very edge of the plot. Their label includes the actual value, so although the "Tar Sands (465)" dot is at about 260 gCO2e/kwh, the actual value is 465 gCO2e/kwh as indicated by the (465).

For any given application, this chart provides a (roughly) optimal ordering of energy sources. Wherever possible, Tidal Range power is the highest EROEI and the lowest (or very very close to the lowest) Carbon Intensity. Next is a choice between Offshore Wind (if you prefer to reduce Carbon Intensity a bit more in exchange for a bit less EROEI) or Sewage / Landfill Gas. And so on.

You'll notice that the included petroleum and petroleum like fuels (US oil, tar sands, Coal) are very dirty, except for possibly "Clean" Coal, which doesn't really exist yet. And there are alternatives to each of them which have higher EROEIs. But none of them have the physical energy density that these carbon based fuels like oil have.

A simple conclusion is obvious: carbon based fuels should not be used for electrical generation (that's what most of these energy sources are, with the exception of Solar Thermal and the generalized cogeneration of electrical power and heat intrinsic to Sewage / Landfill Gas, Concentrating Solar Power, Solid Biomass, and so on) if any of these other alternatives are available and can supply a significant fraction of the required power.

(This post is cross-posted at my personal blog: Mischievous Ramblings II)
There's more...