Maps and charts

I love maps and charts. You will find 22 surprising ones here.

Here are some of my favorites:

Africa is REALLY big:

Compare Canada and Continental Europe:

Colombia’s Data-Driven Fight Against Crime

From the New York Times

"Colombia has always been a violent country, but after the 1948 assassination of Jorge Eliécer Gaitán, a politician who threatened to break the oligarchy’s stranglehold, hundreds of thousands of people were killed over the next five years. The killings were due to scorched-earth battles between the two major political parties — both parties of the upper classes whose major difference was their names. Violence acquired capital letters and became known as La Violencia.

(...)

The dark ages returned with the rise of cocaine in the late 1970s. By the late 1980s and early 1990, the leading cause of death in several of Colombia’s major cities was homicide. The most deadly and infamous was Medellín, but this was also true in Cali.

In 1983, there were 23 homicides per 100,000 people in Cali. Ten years later the rate had reached 120, and some counts put it slightly higher. That’s more than twice as high as the homicide rate ever reached in Detroit, the most dangerous city in any developed country, which peaked in 1993 at 57.6.

But crime in Cali changed after Rodrigo Guerrero became mayor.

What Guerrero did to make Cali safer was remarkable because it worked, and because of the novelty of his strategy. Before becoming mayor, Guerrero was not a politician, but a Harvard-trained epidemiologist who was president of the Universidad del Valle in Cali. He set out to prevent murder the way a doctor prevents disease.

(...)

When Guerrero became mayor in 1992, the conventional wisdom was that the vast bulk of Cali’s murders stemmed from disputes over cocaine trafficking — at the time, the Cali Cartel was overtaking the Medellín Cartel in control of the cocaine trade.

But Guerrero didn’t assume, he measured. The police, courts and every other institution that counted murders all came up with different figures. Guerrero had weekly meetings with these groups and academic researchers to find more accurate figures. Then they mapped homicides by time and neighborhood.

That took about a year — and his term was only two and a half years — but he found something important: deaths were concentrated on weekends, especially payday weekends. 

(...)

“Things that happen on the weekend in our country are often associated with alcohol,” Guerrero said. So Cali started to look at alcohol in the blood of victims (few perpetrators were caught) — and found a large percentage of victims had very high levels. “My initial hypothesis was that this was drug trafficking,” he said. “But the traffickers were not going to wait for weekends to resolve their conflicts — and get their victims drunk.”

The astronomical murder rate was related to the cocaine trade, Guerrero concluded — but only indirectly. Cocaine created social disruption and intensified an already-violent culture. 

(...)

Guerrero banned the sale of alcohol after 1 a.m. on weeknights and 2 a.m. on Fridays and Saturdays. (That 2 a.m. is considered early closing says a lot about the problem.) As he expected, bar owners — and bar patrons — objected. Guerrero asked bars to try it for three months, but success was obvious nearly instantly. The effects were big enough to overcome the objections.

The other decree banned the carrying of guns — enforced by checkpoints and pat-downs — on payday weekends and holidays. The army, which held a monopoly on the manufacture and sale of guns, fought the law. But again, success was persuasive. Researchers compared gun ban days to similar days with no ban in Cali and in Bogotá, which replicated the program. They found that neighborhoods with the ban saw 14 percent fewer homicides in Cali and 13 percent fewer in Bogotá than neighborhoods without restrictions.

Together, those two decrees cut the homicide rate where they were instituted by 35 percent.

There was more: Since the data showed that a large majority of offenders were under 24, Guerrero instituted a curfew for young people in high crime neighborhoods between 11 p.m. and 5 a.m. on weekends.

(...)

The homicide rate declined. From its peak of 120 in 1994, it declined to 80 in 1997. Homicides also dropped in Bogotá and Medellín, which adopted Cali’s program.

Was Guerrero’s program responsible for the drop? It’s hard to separate the effects of the program from outside factors, such as the economy, Colombia’s guerrilla wars and the drug trade (most of the heads of the Cali Cartel were captured in 1995, which could cut either way on homicide). But the reduction in homicide in these three cities was much greater than in the rest of the country.

It’s difficult to say which parts were responsible. There is strong evidence that the gun and alcohol restrictions worked, but the fact that the overall rate in Cali didn’t start to drop until 1994 indicates that other pieces of the program — which were slower to implement and didn’t yield immediate results — mattered as well.

(...)

Perhaps the showcase of Guerrero’s ideas is not Cali, but Bogotá, which never wavered in its implementation of the program. The city’s homicide rate, which was more than 80 per 100,000 in 1993, is 16.7 per 100,000 today.

The gains did not hold in Cali, however. Mayors after Guerrero dismantled almost all the program (two left office early due to accusations of corruption or ties to traffickers). The murder rate went back up again.

But Guerrero is once again mayor of Cali, and once again, the homicide rate is dropping. This year it is on track to drop by more than 30 percent, which would put it at under 60 per 100,000."

When Star Trek meets the APEC summit...

Bayesian statistics applied to Ebola mortality rate

According to this article, the Ebola mortality rate varies a lot across countries:

About two-thirds of the 2,387 people who’d contracted the disease before this year’s outbreak died worldwide. The fatality rate of more than 14,000 people who have been infected with Ebola in West Africa this year is 71 percent. But eight of the nine people who have been treated for the disease in the U.S. have recovered and been released from the hospital. One person has died.


The question addressed in the article is then: how to aggregate this information?

Bayesian statistics seemed like a promising option for estimating U.S. mortality, because it provides a framework for updating prior informed belief (mortality rate in prior outbreaks in the U.S. and elsewhere) with new information (the lower mortality rate in the U.S. in this outbreak).

(...)

Tony O’Hagan, emeritus professor of statistics at the University of Sheffield in the U.K., (...) started with the assumption that the Ebola mortality rate in the U.S. would be 30 percent, about half that in Africa — peppered with a liberal amount of uncertainty because it was essentially an educated guess. Then once he factored in the eight recoveries in nine U.S. cases, his estimate was of a mortality rate of 17 percent.

Waller offered some ideas for how to refine the analysis. What we want is a model that takes into account the individual attributes of each case when estimating the likelihood of death. Those attributes include age and health of the patient, time from first symptoms to start of treatment, training of medical staff and treatments used. What factors led to lower mortality in the U.S., and which can be replicated in the West African countries with climbing caseloads?

Building such a model would require detailed data not just on the nine U.S. patients, but on as many Ebola patients worldwide as possible. That data isn’t always collected and compiled in a usable way, though — especially in an emergency treatment setting.