Wednesday, June 23, 2010

Graphics that Excel 3

The prior post provided an example of 3D Excel objects used to build a large graphic. This post provides five examples of 3D text. Excel2007/2010 3D text functionality opens up design options like shadow, layering, lighting and color gradation. The five examples here use the same text ("TheRaven", of course). Examples demonstrate two simple Excel tricks: (1) you can type in different values for font point size, (2) shadows can be any color. Fonts in these examples are well above the typical 96 point limit found in Excel's font-size dropdown list. The largest shown here is 225 points (images here are smaller than originals). White shadows were used to contrast with dark backgrounds. These examples show a tiny fraction of Excel's graphics functionality.

Tuesday, June 22, 2010

Graphics that Excel 2

The prior post offered a simple image assembled from Excel graphics. This post presents a little more fun in the same vein.

Yeah, I have limited imagination and lack artistic skills. For a three-pound bird with a brain the size of a walnut, I'm Michelangelo. At least I can reach the ceiling without a scaffold.

Graphics that Excel 1

Microsoft Excel is TheRaven's weapon of choice in the war on data. You know data. Not the StarTrek character, who didn't espouse data at all. "Data" in StarTrek was actually the exact opposite - he provided information. Excel is the single best tool for converting mass amounts of useless data into valuable information. Wholesale improvements made to technology underpinning Excel2007 led to a huge increase in capacity. Microsoft also achieved a dramatic improvement in Excel's visual tools, which include charts and graphics. All Excel2007 improvements carry through to Excel2010.

TheRaven constructed a simple image using only Excel graphics. It isn't going to win any awards but it does illustrate some of the flexibility and color options now provided with plain-vanilla Excel. It's theme expresses TheRaven's view regarding a key climate change topic. (Hint: basic chemistry).

If "basic chemistry" didn't make the lights go on, TheRaven doesn't want to be annoying. After all, no one will ever accuse TheRaven of being an artist. The image theme ("ice will melt") refers to inevitability of permanent, consequential climate change in context of a particular compound. Each object in the image represents a methane molecule, composed of one carbon atom and four hydrogen atoms.

Methane is discussed as either as a climate remediation solution or a ticking time bomb. It's posited to replace coal and possibly gasoline. Rising ocean temperatures have sparked fear over possibly catastrophic implications of methane hydrate. Such fears haven't spurred meaningful action. As of 2007, the US got 49% of it's electricity from coal, which is only three percentage points less than in 1990.

Methane in the atmosphere imposes a potent greenhouse effect. It traps multiple orders of magnitude more heat than carbon dioxide. Substitution of methane for coal or uncontrolled release due to suddenly melting of methane hydrates ends up in the same place. Substituting methane for coal would cut CO2 output in half per equivalent unit of energy. This may seem like a step in the right direction but the remaining half is still huge and this solution doesn't address the biggest carbon emitter (China) and the future reigning champ (India).

As China builds coal-fired generating plants like mad and India is desperate to supply 500,000,000 people with electricity, replacing U.S. coal with methane would lower the global price for coal. This would provide a perverse economic incentive to China and India to burn more of it. Given the relentless pace of Chinese and Indian economic growth, the net beneficial impact on global CO2 would likely be negligible. Viewing methane as part of the solution only guarantees that the only ice left on Earth will be indoors and that Mardi Gras will be looking for a new home.

UPDATE: The New York Times reviews a new HBO documentary on domestic gas drilling here.

Saturday, June 19, 2010

Once upon a time in Russia

The Library of Congress website is unquestionably a preeminent Internet treasure. It has collections of documents, posters and photos that span centuries. The LOC website includes thousands of items and covers more than American history.

The LOC maintains a collection of color and B&W photos taken in Russia between 1905 and 1915. From the LOC's website (edited) ...The Sergei Mikhailovich Prokudin-Gorskii Collection features color photographic surveys of the vast Russian Empire. Frequent subjects among the 2,607 distinct images include people, religious architecture, historic sites, industry and agriculture, public works construction, scenes along water and railway transportation routes, and views of villages and cities. The Library of Congress purchased the collection from the photographer's sons in 1948.

From wikipedia (edited) ...Around 1905, Prokudin-Gorsky planned to use technological advances in colour photography to systematically document the Russian Empire. His goal was to educate the schoolchildren of Russia of the vast and diverse history, culture, and modernization of the empire. Outfitted with a specially equipped railroad-car darkroom provided by Tsar Nicholas II and in possession of two permits that granted him access to restricted areas and cooperation from the empire's bureaucracy, Prokudin-Gorsky documented the Russian Empire around 1909 through 1915. He conducted many illustrated lectures of his work. His photographs offer a vivid portrait of a lost world—the Russian Empire on the eve of World War I and the coming Russian Revolution. It has been estimated from Prokudin-Gorsky's personal inventory that he had about 3500 negatives before leaving Russia. Outside the Library of Congress collection, none have yet been found.

The wikipedia article is recommended for its explanation of Prokudin-Gorsky's imaging technique. His color images were stored on multiple plates and were completely reassembled only in the past few years.

TheRaven pulled out a few images from the only complete Prokudin-Gorsky collection on the planet to give you a quick guided tour of early 20th century Russia. Only color images were selected because color seems  much closer to present than monochrome. This is a peek at a Russia yet unaffected by global war. Before revolution, Stalin, starvation, Hitler and the Cold War, when the Tsar still ruled.

The LOC provides remastered versions of some color images. Only the final image in this post is such an image. TheRaven applied light color corrections to other images using Photoshop.Those images are otherwise as provided by the LOC; reassembled from Prokudin-Gorsky's original color plates. Image titles were edited for clarity. Image subtitles were copied from the LOC website without editing.

Studies on Lindozero (laika [dog]) (1915) - Prokudin-Gorskii Collection

City of Lodeinoe Pole, cathedral of Saints Peter and Paul (1909) - Prokudin-Gorskii Collection

Emir of Bukhara (1911) - Prokudin-Gorskii Collection

Evgeniev spring. Borzhom  (1905-1915) - Prokudin-Gorskii Collection

Island Monastery  (1910) - Prokudin-Gorskii Collection
View of the monastery from Svetlitsa [Island, Saint Nil Stolbenskii Monastery, Lake Seliger]

Two men (1905-1915) - Prokudin-Gorskii Collection
Two men standing on a rug, in front of yurt

Three generations (1910) - Prokudin-Gorskii Collection
A.P. Kalganov with son and granddaughter. The last two work in the shops of the Zlatoust plant

Rear platform (1910) - Prokudin-Gorskii Collection
View from the rear platform of the Simskaia Station of the Samara-Zlatoust Railway

Tobolsk (1912) - Prokudin-Gorskii Collection
View of the city of Tobolsk from the north, from the bell tower of the Church of the Transfiguration

Peasant Girls (1909) - Prokudin-Gorskii Collection
Three young women offer berries to visitors to their izba, 
a traditional wooden house, in a rural area along the Sheksna River, 
 near the town of Kirillov.

Group. (Myself with two others, Murman) - (1915) - Prokudin-Gorskii Collection
Sergeĭ Mikhaĭlovich Prokudin-Gorskiĭ (right) 
and two men in Cossak dress(?) seated on the ground.

Thursday, June 17, 2010

False idol

God doesn't approve of false idols, graven images, etc. The bible says so. It is therefore very funny to see God express her wrath with a lightning bolt straight to Touchdown Jesus plastic heart. 'Ole TDJ really didn't hang around very long.  Here a shot taken about eight months before his epic demise.

The Solid Rock Baptist Church erected the half-torso statue in 2004. Nashville artist Brad Coriell built the statue using plastic, foam and fiberglass over a metal frame. The church appears behind Jesus and is located between Cincinnati and Dayton, on the north-bound side of interstate 75. The plastic Jesus was 62 feet tall, measured from the waist. Turns out TheRaven isn't alone in thinking this statue was ridiculous. Controversy and confrontation in the Solid Rock congregation has erupted since its demise.

The next picture shows the depth's of God's wrath. TheRaven is not at all surprised at God's timing, given our real-time oily disaster in the gulf and that most of the Solid Rock Jesus came from oil. Joe Barton better shut his pie hole or God might level the Texas 6th congressional district. If Barton keeps up his corporate kowtow, she'll make the Solid Rock Effigy look like a marshmallow roast.

INRI represents "Iesvs Nazarenvs Rex Ivdaeorvm" and is explained here.

1st picture: TheRaven
2nd picture: Cincinnati Inquirer


Image: New York Times
Caption: TheRaven

What kind?

Image: New York Times
Caption: TheRaven

Friday, June 11, 2010

No record in May (but close)

This is the second post in what will be a series tracking monthly year-over-year changes in atmospheric CO2. Last month set a record but May merely came close, with the 3rd-highest year-over-year increase for the month since initiation of the Muana Loa data series.

The following chart plots minimum and maximum annual changes in CO2 occurring in May, from 1960 to 2010, measuring minimum and maximum values on a rolling 12-month basis. The color area is the spread between minimum and maximum change. Long plateaus indicate periods in which no new records were being set. Choppy areas indicate shorter time intervals between new maximum or minimum values. The white line is 10-year moving average.

Click image to view full-size

Here are the top-3 year-over-year CO2 increases which have occurred in May:

1. May, 2003  +2.96 ppm
2. May, 1998  +2.86 ppm
3. May, 2010  +2.76 ppm

Friday, June 4, 2010

Journalism that reeks

This New York Times article feels good at first blush. The headline news is that Wal-Mart will spend $50 million with an online university, in a new program that supports higher education attainment for rank and file employees. In actuality, Wal-Mart got greater value in free publicity from a credulous New York Times.

American Public University (APU) is headquartered in Charles Town, West Virgina (not to be confused with Charleston), a 1.4 square mile hamlet of less than 3,000 people. APU is Wal-Mart's chosen education partner. Wal-Mart claims that a survey of 32,000 of its employees revealed a strong preference for online education. The Times piece isn't completely amiss. One can deduce that Wal-Mart's low-paid workforce (averaging under $12/hour) will bear the brunt of their education costs. This puts the onus on Wal-Mart to procure cost-effective education for its workforce.

Charles Town is most famous as the place where John Brown was tried and hanged, in 1859. Almost 16% of its population lives below the poverty line. Charles Town has no infrastructure or geographic positioning relative to higher education. A Google Earth view of Charles Town suggests that APU may be located in a residential structure (Google Street View isn't available for Charles Town):

Click image to view full-size
To be fair, APU says its back office operations are located in Manassas, Va. The point on headquarters location pertains to the unlikely confluence of higher-education institutional leadership in a small West Virginia town (that otherwise has nothing to do with higher education).

The times article claims APU has 70,000 students while APU's own website says "...more than 50,000 students in 100 countries.." (a discrepancy of 40%). American Public University is owned by American Public Education, Inc., a NASDAQ listed corporation. The Times article doesn't mention this. From the APEI website: 

American Public Education, Inc. (NASDAQ: APEI) is a leading provider of online post-secondary education focused primarily on serving the military and public service communities. American Public University System (APUS), wholly owned by APEI, comprises two universities: American Military University and American Public University (APU)....APUS serves more than 63,800 adult learners worldwide......In 2002, American Public University was founded.

APU is only eight years old? The Times didn't mention that, either. Based on APEI's financial statements, explosive growth over the past few years has created a cash machine:

  • Revenue grew over 5x from 2005 to 2009.
  • 2009 revenue was $148 million
  • 2009 after-tax margin was 16%
  • 36% return on equity
  • 2010 1st quarter revenue grew 43% over 2009
  • 2010 1st quarter cash flow doubled over 2009
APEI is on-track to become a $200 million corporation this year, generating over $30 million in after-tax profit and about $40 million in free cash flow. If the Times had bothered to read APEI's financials, Wal-Mart's new program wouldn't have been conflated with "wide influence" accruing to a huge employer. Wal-Mart's planned annual tuition assistance budget is only 8% of APEI's probable 2010 revenues. (APEI is also located in Charles Town).

Purely online education has a well-earned reputation for recruiting abuse, student loan fraud and lack of academic rigor. Accreditation and academic worth do not correlate. There are twelve decidedly mixed reviews of American Public University here from current and former students. This sample alone isn't enough to provide a basis for conclusions but the negative comments have an uncomfortable ring of truth.  

University of Phoenix is the big gun of online colleges. It grew explosively in its early years and was subsequently the target of lawsuits and investigations into abuse and fraud.  According to this NPR piece, U of Phoenix anticipated an $80 million fine only last November. Comments posted by current and former students to the NPR piece offer the same disquieting mix of support and derision, in greater volume, with negative comments reinforcing lack of academic rigor as a primary failing. This former U of Phoenix employee has some damning things to say

U of Phoenix is not directly connected to APU but its history, similar business model and APU's stellar growth trajectory beg some questions, such as: how did APU grow so fast? The Times wasn't interested in such inquiry and we are left wondering if APU offers genuine education at a fair price. 
The core problem lies in this statement: "With the work credits and tuition discount, an associate’s degree for a Wal-Mart or Sam’s Club cashier (from APU) would cost about $11,700 and a bachelor’s degree about $24,000". The Times reporter (or editor) lacked the sense to put a $12 thousand Associate's degree in context. It's doubtful that any Times staffers have ever attended a community college. Very few of them likely attended a public university. An $11,700 Associate's degree must look like a bargain to someone who dropped $200,000 on a degree from Columbia

This image points to the problem and perhaps explains some of APU's revenue growth.
APU charges $250 per credit hour, for online-only instruction! Wal-Mart employees will pay 15% less ($212.50), which explains why APU's stated price for a bachelor's degree is $30k while the Times quoted $24k.
Wal-Mart estimates that half of its 1.4 million domestic employees are high-school graduates. Given their average pay and Wal-Mart's modest level of support in the new program, the only education options otherwise within reach for most of these 700,000 potential college students is the nearest community college. Therefore, The Times should have compared APU's pricing to community colleges. Such comparison would determine if Wal-Mart had truly acted in the best interests of its employees.
Community Colleges are a tried and true model, most have decades of experience turning out graduates who are qualified for careers in technology, health care, business roles, law enforcement or for transfer to a bachelors program. Community Colleges offer flexible scheduling and online education options. They are the original evening education model. The President is a believer and this piece in Slate nails the reason: community colleges have succeeded despite a long-running dearth of Federal support. The $12 billion program announced last year works out to roughly $10 million per school. 

Community colleges are also cost-effective, as shown by these examples:
Here's the key comparison:
  • Average community college cost - $85 per credit hour
  • APU's discounted rate for Wal-Mart employees - $212 per credit hour
To a Times staffer sweating a 6-figure student loan debt, the difference between a $12,000 Associate's degree and a $5,000 alternative must appear trivial. To many Wal-Mart employees earning $12 an hour, it's the difference between higher education and hopelessness. Community colleges aren't a perfect solution but, on balance, the education they provide in real classrooms is superior to comparable instruction offered by  "online universities". 

This post identified key facts that The Times missed. If the journalist had invested 1-2 more hours, The Times could have benefited from public interest in an expose of (take your pick):
  • "Wal-Mart's High Price for Employee Education
  • "Wal-Mart Prices Education Out of Employee Reach"
  • "Low-Prices Always? Not for Wal-Mart Employees"
By failing in its journalistic responsibility, The Times became Wal-Mart's unwitting shill. From America's purported newspaper of record, this is journalism that reeks.

UPDATE 1: The Washington Post weighs in here, noting that "For-profit schools may be offering an educational alternative, but that choice often comes with crushing student debt..."

UPDATE 2: Score another win for WalMart. This blogger at theAtlantic gets taken, just like the New York Times.

UPDATE 3: The New York Times published a decent (but small) article that provides online university industry facts & figures, in anticipation of a series of congressional hearings. Senator Tom Harkin has issues a report which states that "profit-making colleges generally charge higher tuition than comparable public colleges, spend a large share of revenue on expenses not related to teaching, experience high dropout rates, and, in some cases, use abusive recruiting and debt management practices." Steven Eisman, a hedge fund manager, has prepared testimony for the hearings and calls the for-profit colleges “marketing machines masquerading as universities.”

UPDATE 4: The Times appears to be catching on to what may be a sensational story of fraud and exploitation, with coverage of a new GAO report. Several undercover operatives documented fraud and deception at 15 for-profit colleges. "...recruiters at four of the colleges encouraged prospective students to lie on their financial aid applications — and all 15 misled potential students about their programs’ cost, quality and duration, or the average salary of in massage therapy and computer-aided drafting that cost $14,000 at a California for-profit college were presented as good values, when the same courses cost $520 at a local community college". The GAO is releasing a video of undercover investigation.

UPDATE 5: The story continues to develop and some ugly outlines are beginning to emerge. This NY Times article provides a key data point:"In the 2008-9 award year, students at for-profit schools represented 26 percent of borrowers — but 43 percent of defaulters. The median federal loan debt for students earning associate degrees at for-profit institutions was $14,000.....Most community college students do not take out student loans". This means that high default rates are driven, at least in part, by "online educators" charging almost 3-times as much for an associates degree as does the average community college. Combined with years of reports of illegal sales tactics and education finance irregularities, this situation has acquired the odor of massive fraud.

Ice will melt

Moral argument will not win the day (or the century) for climate change remediation. Discussion must express change in economic terms; analysis of loss, gain and national competitive advantage. Climate remediation advocates, even thoughtful sites like Real Climate, continue to dwell on science and omit dollar signs. It gets personal with voters before we meaningfully change course. This post provides an example of why, with a look at our dependence on coal for electricity production.

All data used to prepare this post was obtained from the U.S. Energy Information (USEIA).

Fifty-two percent of U.S. electricity produced in 1990 relied on coal. Sixteen years later, coal accounted for 49% of U.S. electricity production. Scant progress in reducing dependence on coal was made smaller still by proportional decline in non-carbon sources. The coal reduction is explained by increased use of natural gas.

 Click image to see full-size
The portion of electricity produced from non-carbon energy actually fell from 1990 (31%) to 2006 (29%). As environmental rhetoric grew increasing strident, dependence on carbon-based energy went in the wrong direction. More than two-thirds of U.S. electricity came from carbon-based energy in 2006.

If we strip away carbon-sourced electricity production, we're left with an interesting fact: reliance on nuclear as a non-carbon energy source increased from 1990 to 2006, from 62% to 67% of all non-carbon sources.

Click image to see full-size

Sources other than nuclear and hydro accounted for 7% of non-carbon production in 1990. This ratio hadn't budged by 2006. The combined nuclear + hydro share of total U.S. electricity production peaked in 1996 and has been on a downward trend ever since. Nuclear electricity production grew 40% between 1990 and 2007, while hydro power fell 17%. While this helps explain stubborn coal dependency, it hardly points the way to clean energy.
Click image to see full-size

Two key points from presentation:
  • 60% of nuclear electricity generation is concentrated in 10 states.
  • Hydro is very concentrated - only four states (Washington, Oregon, California and New York) - account for almost two-thirds of U.S. hydro electricity.
The next chart summarizes 2007 electricity production, by energy source & state. It's sorted from highest-to-lowest level of coal dependency. Thirteen states got at least two-thirds of their electricity from coal, including environmentally-friendly Colorado. Natural gas use is nearly ubiquitous but note gaps in nuclear, hydro and wind power. Some states didn't have access to those types of energy. Coal use is wide-spread. Thirty-six states obtained at least one-quarter of their electricity from coal.

Click image to see full-size

Stubborn coal dependency may be best explained with a question: how do you feel about doubling your electric bill? It may seem like an outrageous question, but recent history points in this direction. As we saw in the prior chart, use of coal to produce electricity is wide-spread but uneven across the U.S. For example, Connecticut in 2007 got only 11% of its electricity from coal, vs. 98% in West Virginia.

Two other questions define further inquiry:

  • How widely do electricity rates vary across the states?
  • Do electricity rates inversely correlate with coal dependency?
We address both questions with the next two presentations, which juxtapose rates and coal dependency in 10 states with the highest rates (left side) with the 10 lowest (right side). Data covers 1999-2007. Rates are expressed as cents per kilowatt-hour (KWH). The rainbow bars report coal dependency.

Click image to see full-size
The two previous presentations show that:
  • Coal dependency in the highest rate states is far below the national average, in all years.
  • Except for outliers that benefit from concentrated availability of hydro (Washington, Oregon, Idaho), the lowest rate states are all heavily dependent on coal. Non-outlier coal dependency is above the national average, in all years.
  • The 10 highest rates exceeded the 10 lowest rates by a range of 50% to more than one order of magnitude (compare blue bars, left to right).
  • Consistent patterns are seen over 1999-2007.
  • There was greater rate inflation in the high-rate states vs. low rate (compare blue bars, vertically), for example, from 1999-2007:
    • California went from $0.1064 to $0.1442, a +36% increase
    • West Virginia went from $0.0627 to $0.0675, an +8% increase
The next presentation drills down into rate inflation and coal dependency by comparing five states with the national average. Prior presentation reported only residential rates. We now look at the four rates classes reported by the USEIA, including residential.

Click image to see full-size

  • Low rate inflation in coal-dependent states over the 18-year period covered by data.
  • The two coal-dependent examples (West Virginia and Missouri) were significantly below national average residential and commercial rates and were also below national average industrial rates after 2000.
  • Examples of high rate, low coal dependency states (New Jersey, California, Connecticut) were way above national averages, in all rate classes, in all years.
  • Examples of high rate, low coal dependency states show significant rate inflation, beginning roughly in 2003, which is not seen in coal-dependent states.


What it means

Coal is the cheapest energy source for producing electricity. Coal fired power plants tend to be older and less expensive to operate. State public utility commissions set rates using return on capital. States with older, depreciated plants (burning cheap fuel) have a capital base that's lower relative to population, so each consumer pays less. The converse is true for high-rate rates, although the mix of technologies and local issues varies. Media focus on cap 'n trade ignores established rate-setting mechanisms that will ratchet up rates according to the level of capital investment needed to replace coal-fired technologies. There's nothing amiss here, it's simply how the system works.

The environmental cause is better served by bracing voters for substantial, perhaps order-of-magnitude increases in their electric bills. Eighteen years recent history show that states in which coal contributes little to electricity production have rates that can exceed 2x those of coal-dependent states. This information isn't clearly predictive, but it strongly suggests that a coal-free future is purchased only at substantial cost to the voter.

The risk of voter rejection is proportional to the pace of change. Public utility commissions designed rate-setting rules in an environment defined by long-lived assets and infrequent technological change. Higher costs of infrequent new plants could be smoothed out against older, depreciated capacity. Potential near-term replacement of one-half to two-thirds U.S. generating capacity, plus new capacity needed by a nation of 400 million people (40 years hence), risks economic shock and populist backlash. The accelerating pace of climate change compels salesmanship, not PowerPoint sermons.

(All analysis by: TheRaven)

Thursday, June 3, 2010

Pond scum, revisited

Least you thought TheRaven harbors personal animus towards Glenn Beck, please read this author interview.

Here's a choice nugget from Alexander Zaitchik, on his new biography of Glenn Beck, 'Common Nonsense'.

Beck has a long history of attacking other people’s kids. In the past, he’s even attacked fetuses. In Phoenix in the mid-1980s, he once called the wife of a rival deejay and mocked her live on the air about her recent miscarriage. In Tampa, he repeatedly called the second-marriage children of Michael Schiavo (husband of Terri) “bastard children.” The tic is just one of the many recurring manifestations of Beck’s molten mean-streak, which runs a mile deep and often explodes like a volcano. 

Where will you be in 2016?

In 335bce Aristotle cranked out Poetics, which has become "...the earliest-surviving work of dramatic theory and the first extant philosophical treatise to focus on literary theory." Who knew that English class is all Aristotle's fault? Fourteen hundred years later, William the Conqueror won the Battle of Hastings. Just shy of 900 years later, Britain maintained economic supremacy in the thick of the Industrial Revolution as the Little Ice Age drew to a close. How much did atmospheric carbon dioxide increase between Aristotle's poetic musings and Europe's great thaw?

Less than two-tenths of one-percent.

Another 108 years transpire between 1850 and NOAA's initiation of CO2 measurement at the Muana Loa Observatory. CO2 increased 12.4% over that period. The Muana Loa data series is now 52 years old. It documents a 22.6% increase in CO2.
  • Aristotle to mid 19th century (2,185 years): +0.18%
  • Mid 19th century to beginning of Muana Loa series (108 years): +12.4%
  • 1958 through 2009 (52 years): +22.6%
The recent CO2 growth curve is close to vertical if viewed in context of western civilization's history.

Two questions:

1. What level of CO2 is considered dangerous enough to merit decisive action?

2. When will such levels be reached?

What level of CO2 is considered dangerous enough to merit decisive action?

This is, obviously, the big question. While there is broad international agreement on existence of an epic problem, progress to-date has been defined mostly by inaction.

350ppm - Dr. James Hansen preceded Al Gore by over a decade as a leading voice for meaningful change. He put his career at risk and some of his early findings, limited at the time by available science, were subsequently proven. Hansen votes for 350ppm as the tolerable maximum level and says "...If we follow business as usual I can't see how west Antarctica could survive a century. We are talking about a sea-level rise of at least a couple of metres this century." The 350ppm goal has been popularized by, which provides a summary of climate science here. The Pew Center maintains a climate advocacy site. RealClimate is the most thoughtful climate science site.

While climate change remediation advocates generally agree on 350ppm as the acceptable maximum level, they almost universally forget to mention close correlation between economics and carbon. Criticism of the U.S. dwells on our 5% of world population producing 20% of CO2. Left unsaid is that we also contribute 20% of global GDP. We crossed the 350ppm level 22 years ago; only six years after China reformed and opened its economy. China's economy is now 12-times larger than it was 22-years ago. Since Wikipedia has a decent overview of economics as climate driver, TheRaven wonders why the remediation advocates don't get it.

400ppm - this is a funny number. Either side of the debate gravitates to 350 or 450, with 400 left to bloggers like this one, who, funnily enough, is also looking at CO2 growth trend. This blogger looks at CO2 from a peak oil perspective and NOAA keeps a running tab.(Note to NOAA and other government agencies - really love the data, you're doing great work, suggest you stop short-selling yourselves with crappy charts.)

The point on 400 ppm is that it's far below long-term targets set by international agreement but well above a comfortable level according to remediation advocates and concerned experts. Hansen completely rejects Europe's 550 ppm target, yet Europe's target is more aggressive than ours. This blog looks at some detail, has better charts that the government and provides footnotes & sources.

450ppm - ClimateProgress wonders if stopping CO2 at 450ppm is politically possible. Hansen's take on 450ppm is that " will probably melt all the ice - that's a sea rise of 75 metres. What we have found is that the target we have all been aiming for is a disaster - a guaranteed disaster...". MIT's Center for Global Change Science developed a climate model that "...looks in great detail at the effects of economic activity coupled with the effects of atmospheric, oceanic and biological systems...". The model was run 400 times. Aggregated results point to high probability of incredible increases in global temperatures. Results were peer-reviewed.

When will such levels be reached?
350ppm - we crossed this line 22 years ago. The decade just ended was that warmest on record. Coincidence?

400ppm - TheRaven extended 52 years of monthly CO2 readings with only a 1% forecast increase in annual CO2 growth. Such assumption is far below historic growth in global GDP. With only a 1% increase in growth rate, 2016 will average more than 400ppm.

450ppm - using the aforementioned 1% increase in growth rates, we cross the global heat Rubicon in 26 years. Considering latent momentum in the $60 trillion global economy, it's simply impossible that CO2 levels will not blast through 450ppm. The mild forecast assumption projects 485ppm by 2050.

 Click image to see full-size

Can it get worse?

2008 was a tough year for the global economy. GDP growth ranged from under 1% for the EU to 9% for China. Global GDP growth was still 3.2% in a tough year. Global GDP growth averaged 4.3% between 1998 and 2003. Accelerating growth in the global middle class may contribute to acceleration of overall growth economic growth. 

If we change the forecast assumption from 1% annual growth in rate of CO2 increase to 2.5% - a level that is significantly below historic global GDP growth - CO2 levels approach Europe's 550ppm goal in a mere 40 years. With the higher growth assumption, we permanently cross 400ppm in 2015 and cruise past 450pm in only 22 years.
 Click image to see full-size
MIT's published results point to as much as a 9 degree (Fahrenheit) increase in average global temperature.

Yes, it can get worse.

UPDATE Aug. 4, 2010: If you live in Russia, it's gotten worse. A lot worse. The Economist weighs in with an insightful analysis of how media is screwing up reporting on climate change.

Crime in Your City

If you want a happy blog, go find TheDove, ThePidgeon or perhaps TheCanary, although anyone adopting a nom de plume of "TheCanary" likely has a few things to say about climate change that aren't terribly happy. This blog belongs to TheRaven. I deal in analysis, insight and commentary. If happy happens, it's likely by chance.

You've read a certain famous poem written by an impoverished 19th drunkard? I was there when he wrote it! Sadly for TheRaven, my contribution was overlooked in a recent NewYorker piece. You'd think that a prize-winning, Harvard historian like Jill Lepore would detect my subtle influence. I gave Poe that whole Lenore shtick. 'Ole weak and weary was rhyming "Sally" with "forgotten lore".  I said "put down the crack pipe" and all Poe heard was "quoth the Raven". One hundred and seventy-five years later, TheRaven shall brighten up your day, nevermore.


We turn to crime - in your city. All data was obtained directly from the Federal Bureau of Investigation. The FBI collects city and county crime statistics under the Uniform Crime Reporting Program and publishes data annually, in September. Be sure to check back for crime insights in 3-4 months, after the 2010 data is available and TheRaven has updated the Crime in Your City reporting tool. This post addresses city crime.

Because the Uniform Crime Reporting Program is voluntary, FBI data represents localities that contain two-thirds of U.S. population. Each annual data set covers three years. Analyses presented here are based on data released in 2009, which covers 2006-2008. The FBI reports numbers of violent crimes, property crimes, uniformed officer staffing and incorporated city population. Data is grouped according to the FBI city classification, which is based on population. "Cities" defined in FBI data are not metro areas but incorporated places. The Census Bureau defines metro areas is a "metropolitan statistical areas", which typically include many incorporated places.

Ten reports are presented in this post. There are only two sets report content. Reports are sorted several ways to highlight changes in police staffing, officer caseload and per-capita crime between 2006 and 2008. Sort keys are indicated by column heading background color. Parameters used to select data from the FBI's 9,400 city/town database appear under report titles. Each report covers a constant selection of 50 cities. National totals and averages appear across the bottom of each report.

If your city doesn't appear, this means that: (1) your city suffers a deficit of crime; (2) your city suffers a deficit of people; (3) your police department suffers a deficit of crime reporting to the FBI. We're focusing on the 50 largest cities in the database that experienced at least two reported violent crimes and at least 18 reported property crimes, per uniformed officer, per year. The FBI data only includes crimes reported to local police.

Very important - caseload figures presented here represent average new cases, per officer, per year, not overall case backlog. The FBI town/city database does not contain backlog or clearance data. The FBI separately reports case clearance data aggregated by city classification. A subsequent post may share this data (but you really won't want to read it).

Report content remains constant in presentations 1-6. A second data-set is provided in presentations 7-10. Presentations 1-6 are oriented around resources available to deal with reported crime. Presentations 7-10 look at crime incidence across city populations.

UPDATE: You can access and use the Crime in Your City tool by downloading it from Docstoc, here. You need zero Excel skills. The tool is 100% point 'n click. It can be used only with Excel2007 or Excel2010. There's also a "Crime in Your City" PDF document which provides a 14-page illustrated walkthrough of the tool, download it here.

1. Police staffing, 2008 - the size of police departments is made comparable by converting uniformed officer staffing figures to per-capita values, i.e., the number of officers per 1,000 inhabitants. The national average is 2.29 officers/1,000 people. The 50 city average reported here is 2.08, roughly 10% under-average. Santa Ana, Ca. is at the low-end with 1.07 officers/1000 while St. Louis has the highest ratio, with 3.94 officers/1000. Eighteen cities are staffed above the 50-city average.

Click image to see full-size

Why is Nashville highlighted in all reports? The tool used to convert FBI data to reportable information (built using Excel2007) produces another report; a single-city crime crime trend analysis. The city selected for trend analysis is automatically highlighted in the Crime Screen report presented here. Nashville happened to be the selected city.

2. Police staffing variance, 2006-2008 - this metric combines two variables in one handy number to indicate relative direction of police staffing. It's based on both city population and the number of uniformed officers. Twenty-five cities increased police staffing between 2006 and 2008 while 25 saw reductions. Changes in staffing levels ranged from a (12%) drop in Toledo to a +10.7% increase in Phoenix. It's important to note that changes in city population can have the greatest impact the police staffing metric.

Click image to see full-size

Why do population numbers "look low"? Because population figures represent inhabitants only within city limits, not the larger metro areas.
3. Violent crime caseload, 2008 - FBI data doesn't identify how many officers are assigned to solving reported violent crimes. The caseload figures presented here are derived by dividing violent crime figures by police staffing. Caseload figures are broadly comparable across all cities. Average violent crime caseload in our 50 selected cities (4.5/year) is slightly more than twice the average in all other localities (2.2/year). Mobile, Al. is the lowest of the selected group with 2.4/year while Stockton, Ca. had 10.5/year. Twenty cities had violent crime caseloads in 2008 that exceeded 2x the national average.

Click image to see full-size

4. Violent crime caseload variance, 2006-2008 - this metric is a rough gauge of changes in police workload, relative to reported violent crime. Between 2006 and 2008, Indianapolis and Bakersfield both saw violent crime increase by at least one reported crime per officer. Nine cities had greater than 10% increases in caseload while 14 cities had at least a 10% decline.

Click image to see full-size

5. Property crime caseload 2008 - same logic that was applied to calculate violent crime caseload (#3, above) is repeated here for property crime. Selected cities average 25.5 reported property crimes per uniformed officer in 2008, compared with 16.4 in all other localities. Detroit was at the low-end with 17.5 while Bakersfield had 46.0. Twenty-five cities exceeded the 50-city average, six of which were at least 2x the national average.

Click image to see full-size

Why aren't these reports more in-depth? Because analysis is limited by available data. There are no technical limitations to refining the analysis to, for example, develop caseload metrics based only of the number of officers assigned to solving violent crimes, or, to study active case backlogs and clearance rates. The limitation is that necessary data is not made available to the public.

6. Property crime caseload variance, 2006-2008 - same logic applied to calculate violent crime caseload variance (#4, above) is applied here to reported property crime. Four-fifths of the selected 50 cities saw a reduction in reported property crime per uniformed officer, between 2006 and 2008. Property crime caseload dropped at least 2x the average of other localities in 17 of the 50 cities.

Click image to see full-size

Report content now changes to look at overall prevalence of crime, beginning with the next presentation.

7. Per-capita violent crime, 2008 - this metric divides city population by reported violent crime to derive the number of violent crimes reported for every 1,000 residents. Per-capita crime is a metric of crime prevalence. St. Louis tops the 50-city list with 20.75 violent crimes per 1,000 residents (which equals to a 2.1% chance individual victimization, per year). Chandler, Az. is at bottom, with 3.17 violent crimes per 1,000 residents. These metrics show that someone living within the St. Louis city limits is 6.5x more likely to be a victim of violent crime, as compared to a resident of Chandler, Arizona. Violent crime is more than twice as prevalent in the 50 cities versus other localities, with 9.45 crimes/1,000 vs. 4.94 in other localities. The prevalence of violent crime was more than 2x the national average in 16 of the 50 cities.

Click image to see full-size

8. Per-capita violent crime variance, 2006-2008 - this metric identifies rising or falling levels of violent crime. Because it combines the number of reported violent crimes and city population, it can be used to compare cities with different levels of population and violent crime. Dallas led the 50-city group with a 25% reduction between 2006 and 2008. Indianapolis was at the other end, with a +25% increase. Twenty-two of the 50 cities beat the national average, which was a 5% reduction.

Click image to see full-size

9. Per-capita property crime, 2008 - this metric replicates logic explained in presentation #7 (above) to measure the relative prevalence of property crime. Reported property crime across the U.S. was 37.9 crimes / 1,000 people in 2006 and fell to 35.6/1000 in 2008, a 6.2% decline. Property crime in the 50 cities averaged 53 crimes/1,000 people in 2008 and all but seven cites exceeded the national average. St. Louis had 85.5 crimes/1,000 people in 2008. While this was a distinct improvement over 2006, it should be noted that St. Louis' 2006 property crime prevalence represented a 1-in-8 chance of victimization per year.
Click image to see full-size

10. Per-capita property crime variance, 2006-2008 - our final presentation replicates #8 (above) with a look at changes in the prevalence of property crime. Two-thirds of the 50 cities reduced reported property crime prevalence to a greater degree than the national average. San Antonio, Atlanta, Glendale and Jacksonville saw property crime increase more than 5%. San Antonio and Atlanta are in the top-5.

Click image to see full-size

What do the reports suggest?

1. Don't blame your police department - caseload figures are a rough gauge of resource allocation.  Police departments dealing with several violent crimes and dozens of property crimes per year, per officer are working hard, with finite resources, to reduce overall crime. Keep in mind,  all crime figures presented here  represent only those crimes reported to the police and caseloads for officers assigned to specific types of crime are much higher than the general averages presented in this post.

2. Crime stats aren't perfectly comparable - one of the reasons why property crime is more prevalent in large cities is that there is simply a lot more property to steal. Municipal governments and police departments aren't the cause of greater concentrations of wealth and property in large cities. On the contrary, this is a factor to consider when allocating municipal resources.

3. Your municipal budget matters (a lot) - relative to crime caseloads,  leverage concepts apply to police staffing.  Nashville's police force numbered 1,230 uniformed officers in 2008, slightly less than 1,250 in 2007. Using Nashville to construct a theoretic example, if one-quarter of it's police force (257 officers) were assigned to investigating violent crimes, then each officer dealt with 33 new cases in 2008.  If  Nashville had increased it's police force by 100 officers at the beginning of 2008, and assigned these officers only to solving violent crime, an 8% increase in overall police staffing would have achieved a 28% reduction in caseload.

4. Above-average crime is a competitive disadvantage - cities compete in more than sports. The arena that matters most is economics. Competition between cities for business relocation and foreign investment is fierce. Investors consider many factors before picking a city for a new factory, distribution center, headquarters or even a branch office. They look infrastructure, workforce quality, average cost of airline tickets, etc. Crime is always on the list. This means that high levels of crime dis-incent new investment, business growth and limit creation of new jobs. Nobody likes taxes but if your city wants to raise taxes  and channel the funds to crime reduction (with oversight in place to make sure this happens), you should support such action.  Other aspects of municipal spending are certainly important but no city can grow and prosper under the burden of high crime.

Finally, all crime reduction programs are not created equal. Ryan Gabrielson and Paul Giblin of the East Valley Tribune won the 2009 Pulitzer Prize for local investigative reporting for their multi-article expose of Sheriff Joe Arpaio. The Pulitzer  committee selected their work  "For their adroit use of limited resources to reveal, in print and online, how a popular sheriff’s focus on immigration enforcement endangered investigation of violent crime and other aspects of public safety." Support your local police department with an eye on effectiveness and, above all, programs that encourage witness cooperation.

You can find the complete Reasonable Doubt series on the Pulitzer website.

Tuesday, June 1, 2010

Is R&D off-shoring killing us?

Richard Florida asks a few questions about R&D off-shoring trends in this post at theAtlantic.

TheRaven ran a quick test to gauge whether R&D off-shoring has caused measurable harm to the U.S. economy. The operative question is whether the economic subset associated with innovation is growing faster or slower than the labor market as a whole. Jobs associated with R&D should grow if the innovation process is out-pacing the economy and compensating for off-shoring. Conversely, if R&D employment is falling, then increased R&D off-shoring can be viewed as an existential threat, not the more benign balancing of national strengths in a global economy.

Bureau of Labor Statistics (BLS) occupation employment data was used to address this question. The BLS publishes pay range and employment data for a constant list of 800 occupations, each May. I have this data by 4-digit NAICS (North America Industry Classification System) code for 2005 through 2009. The BLS publishes data one year after it's initially collected, so, for example, the May 2010 database reports labor market conditions as of May 2009.

There are 47 occupations that can be reasonably associated with R&D - nearly all of the scientific and engineering occupations. Most people in these occupations are not devoted to R&D because R&D is typically just a few percentage points of an overall organization budget. There are exceptions, such as contract research firms, academic labs, etc., but we're dealing with occupation data across the U.S. economy. Looking at these 47 occupations should provide a decent proxy measure, especially given consideration of: (1)  labor market contraction between 2008 and 2009, and that (2) R&D is a well-known budget casualty in tough times.


1. From 2005 through 2009:

a) The 18 science occupations had a +15% aggregate employment gain. Fourteen of 18 occupations had higher employment levels in 2009, vs. 2005.

b) The 29 engineering occupations had a +7% aggregate employment gain. Twenty-three of 29 occupations had higher employment levels in 2009, vs. 2005.

c) All 47 occupations had a +9% aggregate employment gain. Thirty-seven of 47 occupations had higher employment levels in 2009. The overall gain was 164,900 jobs.

2. Between 2008 and 2009:

a) The 18 science occupations had a +1.5% aggregate employment gain. Eleven of 18 occupations had higher employment levels in 2009, vs. 2008.

b) The 29 engineering occupations had a (1%) aggregate employment loss. Sixteen of 29 occupations had higher employment levels in 2009, vs. 2008.

c) All 47 occupations had (0.4%) aggregate employment loss. Twenty-seven of 47 occupations had higher employment levels in 2009, vs. 2008. The net loss was (8,370) jobs.

This group of occupations did better than the labor market as a whole over the 2005-2009 period and especially between 2008 and 2009. BLS data represents the bulk of the US labor market. The NAICS data-sets exclude only farming, domestic help and freelancers.

Click image to view full-size
Conclusion: while the labor market data is broad-brush, employment trends in occupations associated with R&D run counter to adverse macroeconomic circumstance and suggest that increased R&D off-shoring has not materially affected U.S. R&D-associated employment. Therefore, it's not likely that the off-shoring trend discussed in theAtlantic piece has inflicted meaningful harm on the U.S. economy.

PS - this look at employment trends and pay range patterns in four selected science occupations provides further information: