Most Dangerous Cities 2017

More people than ever are migrating to the worlds largest cities for work and play.

However, as the population of urban areas increase it often brings with it an increase in violent crime, exacerbated by high population density and limited law enforcement.

Where are the worlds most dangerous cities?


I used a list of the 50 most dangerous cities (by murder rate) curated on Wikipedia. The data reported is from 2016. Figures include both homicide rate and population by city.

Top 10 most dangerous cities by murder rate

Top 10 most dangerous cities by murder rate (2016)

Download chart.

Homicide danger rank Municipality, Country Continent Homicides 2016 Homicides per 1,000 % Population Killed year
1 Caracas, Venezuela South America 4,308 1.3035 0.13%
2 Acapulco, Mexico North America 918 1.1324 0.11%
3 San Pedro Sula, Honduras North America 845 1.1209 0.11%
4 Distrito Central, Honduras North America 1,027 0.8509 0.09%
5 Ciudad Victoria, Mexico North America 293 0.8467 0.08%
6 Maturín, Venezuela South America 499 0.8421 0.08%
7 San Salvador, El Salvador North America 1,483 0.8339 0.08%
8 Ciudad Guayana, Venezuela South America 727 0.8284 0.08%
9 Valencia, Venezuela South America 1,124 0.7202 0.07%
10 Natal, Brazil South America 1,097 0.6956 0.07%

Full list.

Caracas, Venezuela tops the list of the most dangerous cities with about 1.3 people murdered per 1,000 people (almost 12 homicides per day). Looking at it another way, 0.13% of the population of Caracas is murdered every year.

Acapulco, Mexico (1.13 p/1,000) and San Pedro Sula, Honduras (1.2 p/1,000) are 2nd and 3rd respectively for number of homicides. The top three cities all have a probability of being murdered greater than 1 in 1000!


Download chart.

To put this into perspective, Durban, South Africa, which ranks number 50th for homicides has a murder rate of 0.34 people murdered per 1,000 people (equivalent to 34.4 people per 100,000). Whilst still worryingly high, especially for tourists, it is nowhere near the murder rate of the top 3!

The eagle-eyed readers might have also noticed the top 10 most dangerous cities are all found in either North or South America. No Asian, Oceanic or European cities make the list.

Most dangerous countries


Download chart.

Country Count of cities in top 50 Total homicides 2016
Brazil 19 16123
Venezuela 7 8106
South Africa 3 4190
Mexico 8 3905
Colombia 4 1950
Honduras 2 1872
Guatemala 1 1596
El Salvador 1 1483
United States 4 984
Jamaica 1 545

Full list.

Brazilian cities are deadly! 19 of the top 50 cities by homicide rate are all found in Brazil and account for a total of 16,123 murders (that’s just over 44 per day). Another surprise, for me anyway, was that the United States comes in 4th on the list.

Most dangerous continents


Download chart.

Continent Count of cities in top 50 2016 Total homicides 2016
South America 30 26179
North America 17 10385
Africa 3 4190

Full table.

South American cities make up 30 cities of the top 50. Again, it is important to stress that 19 of these are Brazilian. Brazil would top this list if it was considered on its own. In South America there was 26,179 homicides from just 30 cities in 2016!


I’d love to do a more positive post on the safest cities with fewest homicides (watch your inboxes, subscribers). This analysis would be very interesting¬†to compare to the top 50 most dangerous cities to get an idea of the difference between best and worst cities.

Context is essential to understanding homicide rates. If it was possible to aggregate the data associated with each one of these deaths (from police reports, for example) we could examine the causes of high murder rates.


Caracas, Venezuela tops the list of the most dangerous cities with about 1.3 people murdered per 1,000 people in 2016. Looking at it another way, 0.13% of the population of Caracas was murdered in 2016.

Get the data

  1. Data sources + data used in this post.

Visit Mexico: Sun, Sea… and Homicides

Not many people would choose to holiday in a dangerous area — although some would.

War torn countries often see foreign visitors in the form of armies and press personnel, not those who want to relax on the beach. As such, I believe countries seen as violent attract far fewer tourists than those deemed peaceful (one of the reasons you see Western countries heavily market their safe image to tourists following a terrorist event).

I’ve often heard friends debating¬†visiting countries that I would classify as “safe”. Which left me thinking; should I be worried about visiting certain places that I didn’t immediately considered dangerous? Here are the numbers….


The US State department regularly publishes travel warnings to citizens reflecting dangers within foreign countries. This not only includes warnings not to travel, but also deaths and injuries citizens have suffered in foreign countries. Data.World collated a series of data sets in March 2017 related to these travel warnings for the last 7 years (since 2009 Р2016). In total 240 countries were considered, only 53 of those were issues warnings by the US State Department.


Top 20 Countries by Travel Warnings Issued

Country Region Travel Warnings Issued
Mexico Americas 28
Mali Africa 26
Israel Asia 25
Pakistan Asia 25
Iraq Asia 24
Afghanistan Asia 23
Nigeria Africa 23
Syria Asia 23
Algeria Africa 22
Iran Asia 22
Yemen Asia 22
Burundi Africa 21
Central African Republic Africa 21
Kenya Africa 20
Lebanon Asia 20
Philippines Asia 20
Chad Africa 19
Libya Africa 19
Sudan Africa 19
Colombia Americas 18
Congo Africa 18
Eritrea Africa 18

Full list.

If you¬†expected Iraq or Afghanistan to be top of this list due to recent conflicts, you’d be wrong. Mexico, a country that received 71.5 million US visitors over period considered, had¬†28 travel warnings issued against it. Similarly Israel, a country that welcome 3.1 million tourists in 2015 of which 637,000 we’re by US citizens¬†(4.77 million over period considered), had the third largest number of travel warning issues by the US State Department.

Top Regions by Travel Warnings Issued

Region Travel Warnings Issued
Africa 50
Asia 47
Americas 46
Europe 42
NA 33
Oceania 22

Full list.

The difference between total travel warnings issued by continent is smaller than you may have initially thought. Africa (54 countries) has just 8 more travel warnings than Europe (50 counties).

Note, NA region generally classifies many smaller island territories. See this table for list of countries classified NA.

Top Cause of Death

Cause of death Type Count
Homicide 1243
Suicide 927
Veh. Accid-Auto Vehicle 857
Drowning 776
Other Accident 740
Veh. Accid-Motorcy. Vehicle 247
Air Accident Air 164
Vehicle Accident – Auto Vehicle 152
Disaster 144
Drug-Related 144

Full list.

Homicides are by far the leading cause of deaths for US tourists (more than vehicle accents combined). 593 of these tourist homicides, almost half, were in Mexico! I was surprised to see suicides placed second on the list. Perhaps this is misrepresentation of reported cause of deaths (suicides do not necessarily reflect badly on the country).

Top 10 Countries by Death Count

country region num_deaths num_warnings num_travellers Probability of death
Mexico Americas 1653 28 71608500 0.00231%
Haiti Americas 221 10 3316700 0.00666%
Thailand Asia 191 4 343500 0.05560%
Costa Rica Americas 184 0 8137400 0.00226%
Philippines Asia 181 20 3240000 0.00559%
Dominican Republic Americas 149 0 18169700 0.00082%
Jamaica Americas 128 0 10606600 0.00121%
Germany Europe 125 0 32703500 0.00038%
China Asia 118 0 14593400 0.00081%
Afghanistan Asia 104 23 #N/A #N/A

Full list.

Mexico is looking pretty dangerous based on the death count, but you should pay close attention to the probability of death being relatively low (0.002%, ranked 23rd by probability of death).

Several countries with relatively high rates of American death have not been issued a single travel warning in ~7 years, including Costa Rica, Dominican Republic, and Jamaica. Many of the deaths in these countries are the result of natural disasters where very early warnings are hard to issue by any government.

Top 10 Countries by Probability of Death

country region num_deaths num_warnings num_travellers Probability of death
Thailand Asia 191 4 343500 0.05560%
Pakistan Asia 22 25 226200 0.00973%
Greece Europe 54 0 691800 0.00781%
Haiti Americas 221 10 3316700 0.00666%
Egypt Africa 35 11 579800 0.00604%
Philippines Asia 181 20 3240000 0.00559%
Ukraine Europe 13 15 276800 0.00470%
Saudi Arabia Asia 34 16 749100 0.00454%
Belize Americas 68 0 1563400 0.00435%
Nigeria Africa 32 23 780900 0.00410%

Full list.

The probability (US citizens) of dying in most countries is rarely low, with the exception being Thailand. Here US citizens have a 0.06% chance of death, that is 1 in every 43320 travellers is killed, whilst only 4 travel warning were issued. Pakistan, the country ranked 2nd most dangerous according to State Department data (also ranked 3rd by travel warning issued with 25), has a much lower probability of death, just 0.01%.

Change in travel behaviour based on warnings

travelcountries region num warnings travel pct change
Egypt Africa 11 -34.09
Thailand Asia 4 -15.00
Pakistan Asia 25 -3.00
Philippines Asia 20 -2.34
Venezuela Americas 7 -1.41
Honduras Americas 9 -1.18
Israel Asia 25 -0.06
Mexico Americas 28 0.52
Haiti Americas 10 0.53
El Salvador Americas 9 1.76
Russia Europe 6 4.03
Colombia Americas 18 5.97
Bahrain Asia 3 6.55
Nigeria Africa 23 10.99
Ukraine Europe 15 11.37
Saudi Arabia Asia 16 11.98

Full list.

Overall, countries subject to travel warnings do not see notable declines in American visitors in the 6 months after a warning is issued. Egypt saw the biggest decline in US travellers over the 7 year period between 2009 – 2016, a drop of 34%. Sadly this is the result of political unrest and security fears that have resulted in a significant decline in tourism from most countries around the world.


Political tensions between nations are likely to have a significant effect on the travel warnings issued by countries. Being able to compare the travel warnings issued by each government would be a good extension on this initial analysis (who issues most warnings, where are most warnings issues against, etc…).


The probability (US citizens) of dying most countries is rarely low, with the exception of Thailand. In Thailand US visitors have a 0.06% chance of death.


  1. Data sources + data used in this post.

The Cities Where Breathing is Worse Than Smoking

On hot summer days¬†in London, UK I regularly walk along some of the most trafficked streets (vehicles). Whilst the pollution is¬†not always visible, I can¬†clearly smell the exhaust fumes of the passing busses and feel it settling¬†on my skin. It does not surprise me¬†reading articles with titles including, “London’s Air is as Bad as Smoking” (credited for the title of this post).

Pollution is¬†arguably the biggest hazard to human health today,¬†especially in our cities (although Donald Trump doesn’t agree). With this concern in my mind, I wanted to uncover some of the worst places for my health to visit.

What is PM 2.5 and PM 10?

Particle pollution, also called particulate matter (PM), is a mixture of solids and liquid droplets floating in the air. Some particles are released directly from a specific source, while others form in complicated chemical reactions in the atmosphere.

    • Coarse dust particles (PM10) are 2.5 to 10 micrometers in diameter. Sources include crushing or grinding operations and dust stirred up by vehicles on roads.
    • Fine particles (PM2.5) are 2.5 micrometers in diameter or smaller, and can only be seen with an electron microscope. Fine particles are produced from all types of combustion, including motor vehicles, power plants, residential wood burning, forest fires, agricultural burning, and some industrial processes. PM2.5 particles are considered¬†most damaging to health.


The World Health Organisation produce a Global Urban Ambient Air Pollution Database (data from 2016 update). The primary sources of data for this are official reporting from countries to WHO, official national / subnational reports and national / subnational web sites containing measurements of PM10 or PM2.5 and the relevant national agencies. The dataset covers 2972 cities from 184 countries.


PM2.5 pollution levels by country

PM2_5 Total Country 2014

Interactive map.

Top 10 countries by total PM2.5 pollution level

Country PM2.5 Urban 2014 mean annual¬†őľg/m3 PM2.5 Urban rank 2014 PM2.5 Total Country 2014¬†mean annual őľg/m3 PM2.5 Total Country Rank 2014
Saudi Arabia 127.1 1 107.7 1
Qatar 104.6 2 103.4 2
Egypt 100.6 3 92.8 3
Bangladesh 88.8 4 83.5 4
Kuwait 78.4 7 74.5 5
Cameroon 63.6 12 65.2 6
Mauritania 86.2 5 64.8 7
Nepal 74.3 8 64 8
United Arab Emirates 64.4 11 63.6 9
India 65.7 10 62.4 10

Full table.

Saudi Arabia has levels of PM2.5 pollution in urban areas almost 30% higher than the next most polluted urban areas in Qatar.

For reference, the World Health Organisation recommends¬†10 őľg/m3 (mean annual) as a guideline safe level.

Only 19 of the 179 countries (that reported PM2.5 figures) have equal to or less than the recommended¬†10 őľg/m3 (mean annual)¬†level in urban areas. 20 of the 184 reported countries have total pollution less than this recommendation.

The lowest levels of pollution are found in the Soloman Islands with urban and total PM2.5 levels equal to 5őľg/m3. Notably, larger countries including the United States have urban (8.4őľg/m3) and total (8.2őľg/m3) PM2.5 levels below 10őľg/m3.

PM2.5 pollution levels by city (best and worst)

Rank Country City/Town Annual mean, ug/m3
1 Iran (Islamic Republic of) Zabol 217
2 India Gwalior 176
3 India Allahabad 170
4 Saudi Arabia Riyadh 156
5 Saudi Arabia Al Jubail 152
6 India Patna 149
7 India Raipur 144
8 Cameroon Bamenda 132
9 China Xingtai 128
10 China Baoding 126
2963 United States of America Not in a City,McKenzie,ND 3.20
2964 Canada AUCLAIR 3.00
2964 Canada NORMAN WELLS 3.00
2966 New Zealand Te Anau 2.73
2967 United States of America Not in a City,Custer,SD 2.40
2968 United States of America Not in a City,Converse,WY 2.30
2969 Sweden KIRUNA 2.27
2970 Finland Muonio 2.21
2971 United States of America Wenden, AZ 2.00
2972 United States of America Sinclair,Carbon,WY 1.60

Full list.

Zabol, Iran is the worst polluted city according to the World Health Organisations Air Pollution database by some margin (217 ug/m3 annual mean, 2000% above WHO safe recommended levels!).

4 of the top 10 most polluted cities are in India. 32 of all 2972 cities considered have PM2.5 pollution levels above 10őľg/m3¬†mean annual.

At the other end of the spectrum, many of the least polluted areas are not cities, unsurprisingly. 5 of the least polluted 10 areas are all in the US and all have PM2.5 levels way below the¬†10őľg/m3¬†WHO recommendation.

856 cities considered (of 1972 total) all have less than or equal to¬†10őľg/m3¬†mean annual PM2.5 pollution.


From my brief research it would appear there is some disparity in the way pollution levels are reported. Whilst this analysis provides a good overview to the problem of pollution, it could be worth comparing such analysis to the way data is reported (sensor distributions in countries, etc).

Looking at pollution as a problem, it would also be valuable to compare levels of pollution against population numbers to understand how wide the impact is.


Areas of the Middle East are some of the worst polluted on earth, and most dangerous to your health.

Get the data

  1. Data sources + data used in this post.

British Airways’ IT Failure Cost the Airline ‚ā¨17,246 a Minute

Well, what a weekend it has been for British Airways.

Thankfully I wasn’t travelling out of Heathrow, my local airport, due to the long weekend here in the UK. However I do know of many friends who’ve had their holidays ruined by the “IT failure” that led to the mass cancellation of flights on Saturday and Sunday by British Airways.

The airline has been going through heavy cost-cutting over recent years. I’m almost certain the events this weekend have wiped out any savings they’ve made. Their share price has dropped by ¬£175mm alone.¬†Though how much might¬†this fiasco directly cost the airline?


Airhelp, a company that helps travellers claim compensation for flight delays and cancellations, put together a dataset detailing British Airways flights and estimated passenger numbers for scheduled departures this weekend (27/05/2017 Р28/05/2017) from airports around the world.

In this dataset they split out delayed flights and those that were cancelled. If you booked a flight that departed from Europe or was with a European airline, you  have rights under EU law if your flight is delayed or cancelled. The amount that can be claimed is determined by length of delay, length of flight and wether the flight was cancelled completely.

Using this data it was possible to estimate British Airways’ potential liability for passenger compensation for Saturday (27/05/2017) and Sunday (28/05/2017).

All figure are quoted in Euro’s (EUR).


Worst affected airports (by passengers delayed / cancelled)

Tot affected pax rank Airport Tot delayed pax Tot cancelled pax Tot affected pax
1 LHR 2829 46303 49132
2 LGW 444 4392 4836
3 AMS 284 1153 1437
4 EDI 222 1210 1432
5 JFK 0 1404 1404
6 GLA 243 1072 1315
7 NCE 57 1210 1267
8 GVA 358 847 1205
9 FCO 162 1020 1182
10 DUB 201 910 1111

Full list.

Passengers at LHR (Heathrow), British Airways’ UK hub, saw the worst of the problems. In total 49,132 passengers suffered, 46,303 of them has their flights cancelled.

LGW (Gatwick) was the second worst affected airport for cancelled and delayed flights. 4,836 passengers were affected in total, a much lower number than at Heathrow.

Worst affected airports (by estimated compensation liability)

tot compensation rank Airport 2705 tot compensation (eur) 2805 tot compensation (eur) tot compensation (eur) tot pax
1 LHR 16,819,509 4,414,834 21,234,343 49132
2 LGW 2,067,489 32,324 2,099,813 4836
3 JFK 891,252 891,252 1404
4 DXB 515,516 515,516 668
5 HKG 477,353 477,353 619
6 SFO 435,586 435,586 686
7 EDI 160,470 265,370 425,840 1432
8 AMS 171,197 253,900 425,097 1437
9 GLA 149,743 240,866 390,609 1315
10 GVA 180,207 168,398 348,605 1205

Full list.

Unsurprisingly Heathrow tops the list with a potential liability of¬†‚ā¨21,234,342.77. This equates to ‚ā¨7373 per minute in compensation (21mm / 2880 min)! Ouch.

Remember, one variable that affects the amount of compensation you can claim is the distance of a flight. As BA tend to fly in/out of the UK, passengers flying from/to airports further away from the UK can typically claim more. This is why flights departing from airports¬†like Hong Kong have high total compensation liabilities (‚ā¨477,352.64) in comparison to passenger numbers (619).

Overall, the average compensation¬†liability per airport was¬†‚ā¨245,448.

Worst affected countries (by estimated compensation liability)

Liability rank Country eligible flights eligible seats eligible pax liability (EUR)
1 GB 754 135500 109890 43,490,141
2 US 66 19767 16031 9,618,622
3 FR 78 11322 9182 2,414,753
4 ES 62 9845 7984 2,208,475
5 IT 66 9980 8094 2,187,186
6 DE 59 7876 6387 1,596,859
7 CA 10 2994 2428 1,457,019
8 IN 11 2808 2277 1,366,148
9 GR 24 3840 3114 1,245,696
10 AE 8 2182 1770 1,061,956

Full list.

As British Airways operates out of the UK, it is unsurprising to see compensation liabilities and passengers affected to be highest from¬†British airports. Passengers flying British Airways from the US were the second worst affected (16k passengers /¬†‚ā¨ 9.6mm compensation liability). There were significantly more affected BA passengers from the US than other countries (excluding the UK).

Average claim amounts

Average claim amount by airport histogram

Download chart.

There is a fairly even spread of maximum compensation passengers are eligible to claim.

Typically airports a long distance from the UK – for example DXV, PVG, HKG – have the highest average amount passengers can claim in compensation (‚ā¨771.43) as discussed above.

European airports typically have the lowest amount passengers can claim in compensation (‚ā¨264.53) due to their distance from the UK (or total flight time).

Total liability

tot affected flights tot seats tot pax tot compensation (eur)
Sat 27/5 515 95607 77537 ‚ā¨ 34,193,780.20
Sun 28/5 243 43273 35094 ‚ā¨ 15,476,631.72
Total 758 138880 112631 ‚ā¨ 49,670,411.92

Full table.

In total it is estimated British Airways “IT failure” affected 758 worldwide flights with 112,631 passengers due to fly on them.

Based on EU compensation rules, BA is potentially¬†liable to pay¬†‚ā¨49,670,411.92 in passenger compensation for the two days of delays and cancellations this weekend. Or¬†‚ā¨17,246 a minute (49.5mm¬†/ 2880 min)!


BA is potentially¬†liable to pay¬†‚ā¨49,670,411.92 in passenger compensation for the two days of delays and cancellations this weekend.


  1. Data sources + data used in this post.

The Price of A Big Mac Is Just $1.57 In Ukraine — 75% Cheaper Than In Switzerland

McDonalds has a huge global footprint. How many countries have you visited without one?

Economists have even created¬†the Golden Arches Theory of Conflict Prevention — that is: no two countries that have McDonalds have ever gone to war (the theory has since been disproved).

The Economist has gone so far as creating a Big Mac Index, comparing the price of a Big Mac hamburger around the world.

But which country sells the cheapest Big Mac?


For Big Mac pricing I used The Economists Big Mac Index dataset (1986 Р2016). This includes local cost, exchange rates, dollar price and dollar purchasing-power parity.


Current Big Mac Index (July 2016) – Cheapest vs. Most expensive

Big Mac Index (July 2016)

Download chart.

Rank Country USD Price (Jul 2016)
56 Ukraine 1.57
55 Malaysia 1.99
54 Russia 2.05
3 Sweden 5.23
2 Norway 5.51
1 Switzerland 6.59

Full list.

There’s almost $5 USD difference between the cheapest Big Mac found in Ukraine ($1.57 USD) and the most expensive in Switzerland ($6.59 USD).

The mean average for a Big Mac in July 2016 is $3.61 USD — about the same price of a Big Mac in Pakistan ($3.58 USD, ranked 28/56).

Big Mac Index 1 Year Change (2015 – 2016)

2015 - 2016 Big Mac Index Change

Download chart.

Rank act Rank pc Country 1y actual change USD 15-16 1y % change 15-16
55 56 Mexico -0.74 -31.01%
56 55 Philippines -0.79 -27.88%
52 54 Taiwan -0.40 -18.37%
3 3 India 0.58 24.09%
2 2 Sri Lanka 1.02 28.00%
1 1 Venezuela 2.71 80.20%

Full list.

Venezuela saw a¬†dramatic rise in the price of a Big Mac between 2015 – 2016 ($0.67 – $3.38 respectively) — although I¬†suspect this may be an error in the raw data.

Mexico and the Philippines saw a dramatic fall in the cost of a Big Mac in dollar terms, a drop of around 30% in both countries.

Big Mac Index 10 Year Change (2006 – 2016)

2006 - 2016 Big Mac Index Change

Download chart.

Rank act Rank pc Country 10y actual 06 – 16 10y % 06 – 16
41 41 Norway -1.54 -27.88%
38 40 Taiwan -0.18 -8.32%
39 39 Mexico -0.19 -8.08%
9 3 China 1.48 53.03%
6 2 Thailand 1.84 54.14%
1 1 Uruguay 2.31 56.66%

Full list.

15 countries did not have data for a 10 year comparison (only 41 displayed). For those considered, Norway has seen the largest Big Mac price reduction by dollar value. Uruguay, Thailand, and China have seen the price of a Big Mac increase by over 50% in dollar terms.

Big Mac Index 30 Year Change (1986 – 2016)

1986 - 2016 Big Mac Index Change

Download chart.

Rank act Rank pc Country 30y actual 86 – 16 30y pc 86 – 16
13.00 13.00 Japan 1.07 30.77%
10.00 12.00 France 2.04 45.28%
11.00 11.00 Spain 1.90 49.39%
3.00 3.00 Canada 3.24 70.47%
4.00 2.00 Australia 3.23 75.14%
1.00 1.00 Brazil 4.60 96.24%

Full list.

Over 30 years the Big Mac dollar price has almost doubled in Brazil. In all 13 countries considered the Big Mac has increased in dollar terms. Inflation over this period in the US was 120.2% meaning Big Mac are actually getting cheaper in real terms (in the US anyway).

Big Mac Index PPP (July 2016)

It is important to point out that¬†I used a simple USD price converted from local currency to compare prices in the last three charts. This does not necessarily mean the price of a Big Mac “increased” in local currency. For example, a Big Mac might cost me ¬£1 GBP in 1986 and 2016, but because of fluctuations¬†in exchange rate values the dollar value of the same burgers might vary significantly (e.g $2 in 1986 vs. $6 in 2016).

Calculating purchasing-power parity can show if the Big Mac is undervalued or overvalued compared to the USD (in USA).

Big Mac July 2016 Dollar Valuation

Download chart.

Rank Country Jul2016_dollar_price Jul2016_ppp Jul2016_dollar_valuation
56 Ukraine 1.57 7.7 -68.8
55 Malaysia 1.99 1.6 -60.58
54 Russia 2.05 25.8 -59.32
3 Sweden 5.23 8.9 3.74
2 Norway 5.51 9.3 9.33
1 Switzerland 6.59 1.3 30.79

Full list.

In many countries the Big Mac is undervalued. As an example, anyone travelling from the USA with a pocket full of USD would pay 68.8% less for a Big Mac in Ukraine when compared to home (USA).

In only 4 countries out of 55, is the Big Mac overvalued compared to the USD (USA) — Finland, Sweden, Norway and Switzerland. In Switzerland the Big Mac is over 30% more expensive.


In the interests of fairness, comparing fast-food retailer prices around the world in the same way could make for interesting reading. Does McDonalds huge logistic networks mean they can better compete on price?


The cheapest Big Mac is found in Ukraine ($1.57 USD) — that’s 69% cheaper than buying in the USA.


  1. Data sources + data used in this post.

Airlines Owe Passengers Over 3 Billion Dollars in Compensation (and Counting)

Passengers bumped on overbooked flights has been in the news ALOT recently.

Many flyers on EU airlines (or from EU countries) are becoming privy to the fact they could be eligible to claim compensation for delayed flight in Europe. Airhelp, a company that helps passengers claim compensation from airlines, advises delays over 3 hours are typically eligible for compensation.

If you live in the UK you’ll be aware of the huge liability banks have had for PPI payouts in the last 10 years. With many airlines operating on the brink of bankruptcy according to some sources; how much are airlines potentially liable for in compensation payouts?


Airhelp, a company that helps passengers claim compensation for delayed flights, provided me with a dataset of airports around the world, number of delayed flights (and seats) eligible for compensation, and a calculated estimate of what the total potential liability was for each airport for compensation during 2016 (based on regulation EC261 which covers delayed, cancelled and overbooked flights). Note, compensation is paid by airlines, not airports. Therefore, this post highlights the airports with the highest number of delays, over-bookings, etc, but is not split out to consider individual airlines.

Whilst the dataset covers a large number of airports around the world (5000), only a subset of airports have data for delayed eligibly (574). This is for 2 reasons; 1) compensation rules differ around the world (see EU vs. US, for example) therefore it is difficult to produce accurate estimates for each geography. 2) Airhelphas a commercial interest in the EU market.

Airhelp provided monetary data in Euro’s (EUR). To keep currency consistency with my other posts on this blog I have converted all figures to US Dollars (USD) using the exchange rate of 1 EUR = 1.09 USD (correct 30th April 2017 via

Additional datasets used were from the FAA (US) and CAA (UK) for total aircraft movements at each airport (to calculate the percentage of delayed flights within the US and UK). This data did not differentiate between arriving and departing flights (only departing flights are considered for compensation eligibility), I therefore assumed 50% of operation were departures (roughly in line with LHR figures).


Spread of liability by airport


Download chart.

The majority of airports were liable for less than 10 million in compensation during 2016 (241 airports), though 61 airports were liable for more than this.

241 airports were liable for more than 1 million USD, and 455 liable for over 100K USD, and 566 liable for over 10K USD. Only 9 were liable for less, with VST Stockholm Vasteras Apt bottom of the list with a liability to 14 passengers totalling $3815 USD (avg. $272.50 USD each).

The compensation liability for all airports considered (574) is 3 billion USD ($3,001,339,552.25 USD).

Top 10 airports by liability

rank liability airport code seats eligible for liability liability USD 2016 30Apr17 Seats per flight eligible for liability Avg. Compensation per seat USD 2016 30Apr17
1 LHR 544,343 265,998,150 167 488
2 CDG 370,152 168,216,158 157 455
3 FRA 284,619 118,030,459 143 414
4 MAD 256,543 98,339,909 144 384
5 AMS 250,153 95,679,110 121 383
6 LGW 239,801 87,262,730 143 364
7 JFK 120,174 78,582,024 192 654
8 MUC 235,045 78,321,378 111 334
9 CPH 200,177 70,264,152 121 351
10 BCN 187,368 61,357,899 147 327

Full list.

The top 3 airports (by liability) accounted for 0.5 billion USD of total liability for ll airports ($552,244,766.75 USD). The top 10 airports accounted for 1.1 billion USD of total liability ($1,122,051,967.75 USD), or one-third of total liability.

Airlines who flew¬†out of Heathrow LHR during 2016 were liable for¬†passenger compensation claims to the tune of $0.25 billion USD ($265,998,150.00 USD) — unsurprising given capacity issues at the airport.

Spread of delays

If you booked a flight that departed from Europe or was with a European airline, you might have rights under EU law if your flight is delayed to claim the following compensation:

Delay to your arrival Flight distance Compensation USD 30Apr17
3 hours or more Less than 1,500km 272.50
Between 1,500km and 3,500km 436.00
More than 1,500km and within the EU 436.00
3-4 hours More than 3,500km, between an EU and non-EU airport 327.00
4 hours or more More than 3,500km, between an EU and non-EU airport 654.00

via: Citizens Advice (UK)


Download chart.

Average compensation per seat groups at both end of the spectrum, 175 airports owe an average of $270 – $300 USD per seat (lowest claim possible is $272.50 USD), 174 owe an average of $630 – $660 USD (max claim possible is $654 USD). Note, values are per seat (not passenger) this is because a passenger can be delayed multiple times per year and thus eligible for more than one compensation claim.

Worst performing airports for on-time departures (US /  UK only)

rank liability airport code departures est flights eligible for liability % flights liable seats eligible for liability liability USD 2016 30Apr17
1 LCY 42,585 953 2.24% 55,838 15,336,355
2 LHR 237,482 3262 1.37% 544,343 265,998,150
3 LGW 140,333 1677 1.20% 239,801 87,262,730
4 ABZ 48,078 541 1.13% 23,151 6,319,139
5 SOU 21,412 224 1.05% 8,018 2,205,343
6 EDI 61,110 609 1.00% 48,585 14,382,632
7 GLA 49,064 459 0.94% 24,779 7,058,731
8 BHX 56,592 506 0.89% 41,785 12,705,858
9 LBA 22,152 183 0.83% 7,727 2,348,432
10 NCL 28,132 231 0.82% 11,767 3,738,673

Full list.

UK airports make up all 10 of the worst airports for departure operations by compensation liability. 2.24% of flights taking of from LCY London City Airport were liable for compensation claims from passengers.

Note, this dataset only contains US and UK airports for comparison.


There are two clear improvements that can be made to allow for a more comprehensive analysis.

Firstly, having access to addition compensation data for more airports with local compensation rules (vs. just EU) could provide interesting analysis into best regions for compensation, worst regions for liability, etc.

Secondly, accurate operations departure data for all major worldwide airports would improve calculations. Due to the datasets used, the above workings only consider some US airports and UK airports, whilst also use estimated commercial departure figures (departures=total commercial operations / 2).


Airlines who flew out of Heathrow LHR during 2016 were liable for passenger compensation claims to the tune of $0.25 billion USD ($265,998,150.00 USD). The compensation liability for all airports considered (574) is 3 billion USD ($3,001,339,552.25 USD).


  1. Data sources + data used in this post.

We Are Now Taking Off From 22000 Feet

It is very likely you’re within 50km of an airport — even if it’s a very small one.

Passenger volumes continue to rise year-on-year, and newer, larger airports are needed to keep up with demand. My local major airport, London Heathrow (LHR), claims to operate at 98% capacity.

London alone has 5 major airports, and many¬†fast growing commercial ones such as London Southend (SEN). Although if¬†it’s actually in London is debatable!

With a recent inquisitiveness about airports, I began questioning. How many airports are out there? Where are they? Here are some interesting statistics.


OurAirports maintain a comprehensive document of airports around the world. It currently lists over 46000 airports and includes information such as location, elevation, runways, and identification codes. This data was used to perform all analysis.


Airports by country

Rank by airport count country count_airports
1 United States 21487
2 Brazil 3833
3 Canada 2434
4 Australia 1879
5 Russia 919
6 France 789
7 Argentina 710
8 Colombia 701
9 Germany 663
10 Venezuela 592

Full list.

Perhaps to be expected the US has the most airports, but I did not expect it to be almost 18,000 more than the second placed country, Brazil. Although most of the airports in the US are classified a small (private) airports, 21.5k airports is still a huge number!

Worlds highest / lowest airports

elevation_rank type name country elevation_ft elevation_mt
1 heliport Siachen Glacier AFS Airport India 22000 6,706
2 small_airport Daulat Beg Oldi Advanced Landing Ground India 16200 4,938
3 small_airport Laguna Choclococha Airport Peru 14965 4,561
4 small_airport Fausa Airport Peru 14809 4,514
5 small_airport San Rafael Airport Peru 14422 4,396
46231 small_airport Brawley Municipal Airport United States -128 -39
46232 small_airport Ein Yahav Airfield Israel -164 -50
46233 small_airport Cliff Hatfield Memorial Airport United States -182 -55
46234 small_airport Furnace Creek Airport United States -210 -64
46235 medium_airport Bar Yehuda Airfield Israel -1266 -386

Full list.

To give this some perspective, Mount Everest stands at¬†8848¬†meters¬†— just 2000 meters¬†higher than Siachen Glacier AFS Airport, India. In fact, this airport (heliport) would rank just outside the top 100 highest mountains. Sadly, it only serves a military purpose, but all credit to the pilots flying at that elevation! The highest large airport is Alejandro Velasco Astete International Airport in Peru at 3310¬†meters above sea level.

At the other end of the scale, Bar Yehuda Airfield in Israel lies 386 meters below sea level. The English Channel is only 120m deep.

Worlds longest runways

runway length rank type name country length_ft length_mt surface
1 seaplane_base Pontiac Airpark Water Aerodrome Canada 120,000 36,576 water
2 seaplane_base Gunflint Seaplane Base United States 30,000 9,144 water
3 seaplane_base Libby Camps Seaplane Base United States 26,000 7,925 water
4 seaplane_base Brookville Reservoir Seaplane Base United States 25,000 7,620 water
5 seaplane_base Long Lake Seaplane Base United States 25,000 7,620 water
6 seaplane_base Conchas Lake Seaplane Base United States 21,120 6,427 water
7 medium_airport Qamdo Bangda Airport China, People’s Republic of 18,045 5,500 concrete
8 seaplane_base Goddard Seadrome Seaplane Base United States 17,000 5,182 water
9 medium_airport Ulyanovsk East Airport Russia 16,404 5,000 concrete
10 medium_airport Pierre Van Ryneveld Airport South Africa 16,076 4,900 asphalt

Full list.

No that’s not a typo,¬†Pontiac Airpark Water Aerodrome has a runway that’s 36.5 km long (it is water runway for seaplanes). The longest traditional runway is at¬†Qamdo Bangda Airport in China at 5.5km. The runway needs to be so¬†long because of the¬†elevation of the airport (4333m). When the air is thinner (at altitude) it takes longer for a plane to take-off and land.

The runway at Juancho E. Yrausquin Airport in Saba (in the Caribbean) runs a mere¬†396m, making it the world’s shortest commercial runways (excluding heliports).


I would be very interested in analysing how much traffic some of the runways with notable characteristics receive and any associated accident information, especially those with a high elevation.


The United States has 21.4k airports. That’s almost 18k more than any other country.


  1. Data sources + data used in this post.

Concorde Would Still Beat Hyperloop

High-speed trains now compete with airlines for short-haul routes.

The fastest high-speed train in operation is limited by a maximum speed of 431 km/h, although trains have reached speeds of 603 km/h in testing.

Hyperloop is a proposed new mode of passenger and freight transportation that propels a pod-like vehicle through a near-vacuum tube at airline speeds. Top speeds of 1200 km/h have been touted. To put that in perspective a Boeing 777-200-LR, one of the fastest commercial jets in operation has a maximum operation speed of 1037 km/h.

What could a world with interconnected Hyperloop routes look like?


Given Hyperloop is still in development phase, with some limited testing completed, much of this post takes into account assumptions about what is possible (speeds, geography, engineering limitations etc).

Many of the routes I’ve considered¬†are already in operation — served either by plane or train. Distances for these routes via Hyperloop are very rough given I have not considered geography in any detail. That said,¬†because of the physics involved Hyperloop needs a very straight track.

The aim of this post is to provide a basic overview of what could be made possible with Hyperloop.


Maximum operating speeds

Max operating speed by mode of transport

Download chart.

The fastest of each mode of transport currently in operation are considered: Hyperloop, Boeing 777-200LR (plane), Shanghai Maglev (train).

Hyperloop is most similar to the existing train networks. The difference between current high-speed trains and Hyperloop is massive (1200 km/h vs. 431 km/h — almost 3 times faster.).

Average Speeds for Popular Air / Rail Routes

Plane vs train vs Hyperloop

Download chart.

For the above 5 routes, all of which show actual times for plane and train journeys, the train is always the slowest (almost twice as slow as a plane for most routes).

Travelling by plane would be¬†the second quickest option, with Hyperloop coming a clear first for all routes (in some cases Hyperloop is almost 4 times quicker on some routes — Paris to Lyon, and twice as quick as taking the plane on others).

It is important to note these figures only consider time spent moving, excluding check-in time, etc.

There are a number of cities currently considering implementing Hyperloop. LA to San Francisco, Dubai to Abu Dhabi, Helsinki to Stockholm, etc. Most of these routes are well below 1000km, as are the routes considered above with the exception of Beijing to Shanghai at just over 1000km apart. At an estimated $19.1 million USD per kilometre, cost is the biggest inhibiting factor to a longer Hyperloop routes at present.

Did you know? It will cost an estimated $2.65 billion USD ($19.1MM USD x 139km) to build a Hyperloop route between Abu Dhabi and Dubai.

Hyperloop potential

Concorde once circumnavigated the globe in 32 hours 49 minutes and 3 seconds, starting and ending its journey in New York, via Toulouse, Dubai, Bangkok, Guam, Honolulu, Acapulco, to refuel and then back to New York JFK (a total of 36,787.6 km).

Assuming Hyperloop took the same journey at its average estimated speed of 970 km/h (which unlike a plane would not need to account for refuelling time), it would take almost 38 hours (36787.6 km/970 km/h). That’s 5 hours slower than Concorde.

Hyperloop round the world

Full map.

The circumference of Earth at the equator is about 24,874 miles (40,030 km), but from pole-to-pole ‚ÄĒ the meridional circumference ‚ÄĒ Earth is only 24,860 miles (40,008 km) around. This shape, caused by the flattening at the poles, is called an oblate spheroid. A “true” round-the-world journey circling¬†the equator would take over 41.3 hours (40030 km¬†/970 km/h) at a cost to build of $764.6 billion USD (40030 km¬†* $19.1 million USD per km).


As discussed, many of these calculations are very rough. When Hyperloop gets closer to a production version where better estimates are available the calculations used can be improved.


Concorde once circumnavigated the globe in 32 hours 49 minutes and 3 seconds, the same route on Hyperloop would take almost 38 hours.

Get the Data


1 Litre of Beer Costs Just 1 USD in Slovakia

Economists use a range of goods to gauge consumer prices in different countries around the world.

Beer is one consumable that can be used for such purposes, although not often omitted.

I believe this is because unlike a bottle of Coca-Cola, the local varieties of beer can result in significant price differences. Which left me wondering; where can you visit to find the cheapest beer?


I used GoEuro’s Beer Price Index (2015 and 2016) for beer costs around the world. This index includes both supermarket and bar costs for a 33cl beer. GoEuro’s data¬†also calculates a mean average beer price¬†using the calculation (supermarket cost + bar cost) / 2.


Cost of beer around the world 2016

Overall Average Price for Beer in USD 2016 Histogram

Download chart.

A 33cl beer will cost you anywhere between $1.14 and $6.59, on average. Lausanne, Switzerland is the most expensive where a beer at a bar will cost you $12.20. That’s over 6 times more expensive than the cheapest city for a beer at a bar, Kiev ($1.82).

Cheapest beers 2016

Beer Price Index 2016 map

Interactive map.

Overall rank (overall ave) City Country Average Price for 33cl Beer (Supermarket) in USD Average Price for 33cl Beer (Bar) in USD Overall Average Price in USD Bar – supermarket diff Bar – supermarket diff pc
1 Bratislava Slovakia 0.35 1.93 1.14 1.58 451.43%
2 Kiev Ukraine 0.47 1.82 1.15 1.35 287.23%
3 Cape Town South Africa 0.46 2.12 1.29 1.66 360.87%
4 Krakow Poland 0.57 2.02 1.30 1.45 254.39%
5 Mexico City Mexico 0.46 2.48 1.47 2.02 439.13%
6 Belgrade Serbia 0.30 2.73 1.52 2.43 810.00%
7 Cairo Egypt 0.69 2.61 1.65 1.92 278.26%
8 Warsaw Poland 0.60 2.79 1.70 2.19 365.00%
8 Ljubljana Slovenia 0.75 2.64 1.70 1.89 252.00%
10 Budapest Hungary 0.28 3.13 1.71 2.85 1017.86%

Full list.

Most of the cities in the top 10 (7) are all found in Eastern Europe. In Bratislava, Slovakia a 33cl beer can be bought for $0.35. At a bar in Bratislava the same beer might cost you $1.58 — a 451% increase! In Budapest the difference between supermarket ($0.28) and bar ($3.13) is 1018% (Paris is the worst — a 1270% increase).

Year-on-year beer price changes

2015 - 2016 beer price change

Download chart.

Price change rank City Country Overall Average Price % Change 15/16
1 Rio de Janeiro Brazil -43.27%
2 Geneva Switzerland -42.80%
3 Munich Germany -42.05%
4 Cape Town South Africa -39.72%
5 Bucharest Romania -37.81%
6 Ljubljana Slovenia -37.34%
7 Tel Aviv Israel -37.10%
8 Dublin Ireland -34.32%
9 San Francisco United States -34.01%
10 Barcelona Spain -33.90%

Full list.

GoEuro’s price index records beer prices falling in 63 of 64 cities between 2015 and 2016! Only Malaga in Spain saw beer prices increase, by 23% — sorry packaged holiday makers.

The fall in beer prices was actually fairly significant with an average price decrease of 24%. The top 10 cities ranked by falling beer price all saw price decreases of over 30%. In Rio De Janeiro, Geneva, and Munich average beer prices fell by over 40%! Oktoberfest is looking cheap in 2017!

Beer consumption 2016

Beer consumption 2016 map

Interactive map.

Did you know? Out of the 70 countries considered in GoEuro’s Beer Price Index, the top beer consumers in 2016 were: 1) Czech Republic (144 litres person / year), 2) Poland (127), 3) Serbia (119).

Biggest spenders on beer 2016

Average spend on beer consumption 2016

Download chart.

Av spend rank City Country Av. Annual Consumption Per Capita (litres) Av. annual spend per capita (Supermarket only) Av. annual spend per capita (Bar only) Av. annual spend per capita (average)
1 Lausanne Switzerland 56 166.30 2,070.30 1,118.30
2 Helsinki Finland 105 381.82 1,753.18 1,067.50
3 Sydney Australia 99 540.00 1,398.00 969.00
4 New York United States 85 265.30 1,645.91 955.61
5 Miami United States 85 182.88 1,602.12 892.50
6 Toronto Canada 116 376.12 1,395.52 885.82
7 Prague Czech Republic 144 318.55 1,444.36 881.45
8 Venice Italy 106 212.00 1,532.18 872.09
9 Moscow Russia 106 395.09 1,323.39 859.24
10 Frankfurt Germany 107 220.48 1,420.18 820.33

Full list.

Whilst Czech Republic, Poland, and Serbia have the highest beer consumption per capita, the relatively cheap price of beer means that they don’t spend as much ($881 per year, $652 and $546 respectively).

Unsurprisingly Lausanne, Switzerland Рwhere a beer at a bar can cost you more than $12 Рcomes out top for most spent per person on beer each year at $1,118 (even though consumption is relatively low at only 56 litres p/person p/year).


GoEuros dataset has aggregated beer prices. Having access to the raw dataset with prices by brand would be particularly useful in comparing like-for-like differences in prices around the world.


33cl of beer is cheapest in Bratislava, Slovakia at an average cost of $1.14 ($0.35 in a supermarket or $1.93 in a bar).


  1. Data sources + data used in this post.

There Is Almost 2 Trillion USD of Planes in Our Skies

$1,911,494,699,999 USD to be exact.

I don’t need to tell you jet planes are expensive. Very expensive. Spoiler alert: you (probably) can’t afford one.

That said, commercial airliners are abundant as a quick look on Flightradar24 will confirm.

At hundreds-of-millions a piece; what is the value of the current fleet of commercial airliners in the sky?


I used the average cost of a commercial jet model as supplied by the manufacturers who dominate the commercial market, Airbus and Boeing.

Fleet data was obtained from AirFleets for the SkyTrax top 100 airlines (2015).

Total price values were calculated using average cost of model and most popular model variation in operation. For example, count of the more popular A350-800 was used over the A350-1000 variant.

Aircraft data (seat numbers) was obtained from Wikipedia.


Most expensive planes


Download chart.

Unsurprisingly the massive Airbus A380 is the most expensive commercial airliner. All of the top 10 planes by price are worth more than $300 million USD each.

Most expensive cost per seat


Download chart.

The A380 represents represents a high cost efficiency per seat (rank 15th @ 0.82$ million USD / seat).

The A350-1000, 777-9, and 777-200LR all have per seat values of over $1 million USD (cost / number of seats). This is a rough calculation assuming the maximum number of seats are installed (in many cases this does not happen).

Total value of planes by type


Download chart.

Did you know? The cost of planes for all of SkyTraxs top 100 airlines (new) is $1,911,494,699,999 USD. 

The Boeing 777 ($340 million / each), 1214 in operation , have a cumulative new value of $412,274,400,000 USD (412 USD billion).

The cheaper 737s ($110 million / each), of which there are 3471 planes in the sky (25% of all commercial aircraft) (almost 3x the 777), have a total combined value of $381,810,000,000 USD (382 USD billion). Southwest operates 710 737s. Bought new these would cost $78,100,000,000 USD (78 USD billion).

Boeing versus Airbus

Boeing Airbus total value

Download chart.

Although Airbus offers the most expensive plane for sale (the A380 @ $432 million USD), the new cost of all Boeing planes in the sky is much higher than Airbus.

There are over $1 trillion USD of planes flying from Boeing (one trillion sixty-eight billion seven hundred thirty-one million seven hundred thousand), whilst Airbus has just $800 billion USD (eight hundred forty-two billion seven hundred sixty-three million).


All the calculations and variable values (plane cost, seats on aircraft, ticket cost, etc) are nowhere near accurate and should only be taken as a generalisation.

Should true data be obtainable, values for these variables would greatly improve the accuracy of all data and calculations used in this post.


The cost of planes for all of SkyTraxs top 100 airline (new) is $1,911,494,699,999 USD. 


  1. Data sources + data used in this post.