Visualising the Guardian Datablog

Friday, November 6th, 2009

I’m doing a regular weekly visualisation for the excellent Guardian Datablog, the front-end for an amazing library of statistics and data, lovingly hand-gathered by The Guardian.

My first post is about Deadly Drugs.

There’s been a furore over here in the UK about the dangers of illegal drugs. The Government has sacked its most senior drugs advisor, Dr Professor Nutt, after he claimed cannabis was no more harmful than alcohol. And that horse-riding, and specifically ‘equasy’ (Equine Addiction Syndrome) was riskier than taking ecstasy. (Statistically he’s correct. His study here.).

Anyway, digging at the numbers behind his statements and how drugs are reported in the popular press, I found some stuff I didn’t expect about drug harms.

Check out the article on The Guardian blog for detail and data. You want both right?

Books and Store

Our Beautiful Books - Information is Beautiful Information is Beautiful Store

Show Comments ( )

  • Jason

    I really like to see informations this way, it’s brillant!

    Is there a mistake with “solvents”? 12 deaths, 10 press coverage … should be 83%, not 126% =)

    Keep the good work!!

  • Michael

    No alcohol or tobacco statistics? I think it’d be helpful if they were included.

  • Esme

    The disproportionally higher reportage of cannabis, ecstasy and cocaine is expected. Though seeing a similar pattern for aspirin and solvents isn’t.

    The biggest surprise to me was the sheer number of deaths from paracetamol and non-SSRI antidepressants. The UK label for paracetamol is much more strongly worded than in the US, and it wasn’t until I was trying to find some infant Tylenol equivalent in the UK that I knew it had complications like respiratory and kidney failure.

  • ismael peña-lópez

    Dear David,

    Thanks for the blog, which I enjoy with every post.

    I think that some copy-and-paste accident happened in your last figure picturing the “Deadliest Drugs”. In the last bubbles (“solvents”) there’s a 126% when it seems, by the figures (12 – 10) that it should be 83%.

    Can you clarify that? :)

    Thank you very much,


  • Simon Miller

    Hi, is very cool.

    Just querying the figures for solvents – 12 deaths and 10 reports, shouldn’t this be less than 100%? Or am I missing something


  • Dylan

    Shouldn’t that last one be 83%…?

  • teedyay

    Typo on the last line, surely? (10/12=126%?)

  • teedyay

    … sorry: otherwise, I love it!

  • martin

    10 is not 126% of 12.

    • david

      Thanks all for pointing out the typo. 10 out of 12 is 83%. Fixed now.

  • Joseph Rushton

    On your Guardian article you posted the second image (Deadliest Drugs II) twice and forgot the first image (Deadliest Drugs).


  • Janti

    Another great post.

    David, How do you draw these diagrams? Is there some “special” graphics plotting software that I haven’t heard about, or do you labor through Illustrator et al?

    • david

      @janti all lovingly hand-tooled in Illustrator.

  • PaulBo

    Do you have any further information on the antidepressant and paracetamol deaths?

    I’m just wondering if these figures are skewed by suicides and don’t represent the risks from taking the recommended dosage of the drugs themselves.

  • Jessica

    Would be interesting to organize the data by the right-side column, i.e. drugs ranked by what is reported on the most, to see how many deaths are actually caused by societally-unaccepted drugs vs. more-accepted drugs (i.e. cocaine, #1 reported on w/ 235 deaths vs. alchohol, #9th reported on w/ 685 deaths. Or, alcohol reports are only 8% of cocaine reports while alcohol deaths are 290% more than cocaine deaths)

  • swag

    Go solvents.

  • Hayles

    As others have said, this is great information, clearly presented. I’d love to see a follow up comparing the number of doses taken, pills popped, caps rolled, etc. (all estimates of course) to the deaths. Does the Guardian curate that sort of info?

  • Web Architecture

    Surely to call something more or less deadly, the ratio of users to deaths is a critical factor?

    Let’s say Drug A has 10 users and they all die — it’s 100% deadly.

    And let’s say Drug B has a 1,000 users and 100 die… it’s 1% deadly, even though 10 x more people died of drug B than drug A.

    Without figuring in the # of users, this chart is totally misleading!

  • Mick

    Can I ask where you got the statistics for Cannabis deaths? I have read many different sources that say there has never been 1 single death attributed to using the plant in recorded history (not saying there hasn’t been any deaths while ‘high’). I think the ‘highly questionable’ sums it up though ;)

  • gifts for men

    I think, the visualization turned out really cool and also useful.And the world map view is not scalable. I suppose making this dynamic would take care of the information overload… by allowing one to choose which regions you wanted to compare.Anwyas, i thought it was pretty cool.

  • Alex98765

    Cannabis deaths come from car accidents I’m sure… people who forget which pedal is which and people who wait for a stop sign to turn green… u know what i mean. People high on cocaine and only cocaine are typically too afraid to even drive, but sometimes they have to. Their deaths are from OD’s usually combined with other drugs. Drunks will drive anywhere, anytime….

  • Brendan

    I would like yo know where the stats have come from to show that cannabis has killed ANYONE other than maybe having been the cause of a fire that may have killed someone. I did notice at the bottom of the diagram there is “*cannabis statistics highly questionable” but you could have safely put 0 and had sufficient stats to back up that number if it were to be refuted.

  • Coin

    uhh so what are the cases where someones died from marijuana?

  • Chris

    Clearly, some of you all still didn’t understand the data. The percentages were that of deaths reported. In cases where there were more reports than deaths, you should clearly get a percentage larger than 100 as more than the actual deaths were reported, either through multiple sources or the same death being reported more than once (this all goes into how the data was acquired however).

    10/12 = 83%



    and so on.

  • aljuk

    Very interesting, thank you.
    I’d question the ecstacy figure – I wonder how many of those fatalities are down to water toxicity, and re. cannabis “fatalities” I have yet to hear of even one case in recorded history of anybody overdosing or dying as a direct result of using cannabis, so I would imagine those fatalities must be attributable to driving accidents etc.

  • Neuroskeptic

    Fascinating, it seems like you’ve independently replicated a previous study (which David Nutt has quoted several times) from way back in 2001 – Distorted? a quantitative exploration of drug fatality reports in the popular press

    Like you he found 100% coverage of ecstasy deaths and fairly high coverage of heroin deaths, but very low coverage of alcohol and OTC drugs.

  • Steven

    No 1 has evr died from smokin cannibus, only 2 deaths have occured and those 2 choked on their own vomit, not the weed!! This blog is wrong

  • Lincoln H

    120% of this is made up. it is almost impossible to overdose on cannabis, it requires inhalation of over 1k pounds.

  • Dave

    And people, do note that not only are the cannabis numbers “highly questionable” but it has been proven many times that it is virtually impossible to die from a thc overdose. Read here for example: .

  • Dave

    And people, do note that not only are the cannabis numbers “highly questionable” but it has been proven many times that it is virtually impossible to die from a thc overdose. Read here for example: Wikipedia.

  • Apotheote

    No one in the entire history of the universe has died as a direct result of marijuana, so the % should be infinity. I have no idea where you found evidence for 19 deaths a year…

  • MikeF

    Are you saying that 19 people died from actully smoking cannabis? Or died from something else trying to obtain it?

  • Koz

    Can anyone clear up for me how the death’s from cannabis occured in this report? I always believed it to be non lethal (directly at least).

  • Chris

    This is crap there has been no such thing as Cannabis Intoxication

  • The Heretic


  • ahrcanum

    No dealths associated vaccines or does that fall under solvents?

  • jon

    Congratulations, David! Great news!

  • B

    To be fair, it seems like drunk driving fatalities could be included with “alcohol deaths”. But, if we’re allowing that, then we should allow for drug-related deaths that aren’t due to overdose (say, in the case of marijuana). So, everyone pointing out “you can’t overdose from marijuana” should stop and think about how drug-related fatalities are being counted.

    Also, as someone pointed out, there are no tobacco deaths on the list. While no one has overdosed from it, I’m sure tobacco-related cancer dwarfs all the other deaths on the list combined.

  • speed dating london

    I’m doing a regular weekly visualisation for the excellent Guardian Datablog, the front-end for an amazing library of statistics and data, lovingly hand-gathered..

  • nicola

    surely some of these are accidental and some are deliberate. I can’t see somone accidently dying from paracetemol or aspirin or am i wrong? does this cover both cases?

    • david

      @nicola I think accident overdoses of paracetamol do happen, but they are the rarity.

  • Trevor

    this information is wrong. there are hundreds of studies showing its impossible to die from cannibus

  • Michael

    David, this really is a beautiful visualization.

    A suggestion (and request!). Colour the balls according to the drug’s legal class, and make 3 graphs: (1) sorted by risk of death; (2) sorted by publicity; (3) sorted first by legal class, then by publicity.

  • Tsahi Levent-Levi

    Statistics – so easy to fool around with.
    Alcohol kills more simply because more people consume it. There is no indication here as to the population size of the consumers of each, so the numbers start skewed as it is.
    I am sure that more people in China die from solvents than people die in the UK from alcohol – it’s just the difference in population sizes.
    The visualization is cool and teaches about publicity of deaths of each, but not about which is deadlier.

  • Jim

    The cannabis deaths are indeed highly questionable. When it comes to smoking/ingesting plants, there are a lot of variables to consider (i.e. soil/plant composition, etc.) It’s really hard to prove that THC or some other active cannabinoid(s) are the cause of death.

  • Feldur

    Don’t be silly, people, there are many ways cannabis can kill. For instance, one could be hit over the head with a slab of hash. Or drown in a vat of cannabis oil. Or be crushed under precariously stacked piles of Mexican brickweed, etc, etc…

  • http://Guardiandatablogondeadliestdrugs Vicky McAndrew

    Another good example of press sensationalism. To actually report more deaths than have occurred is outright lying. Is there nothing that can be done about their swaying of public opinion with statistics that are made up? It has always seemed to me that the public have sheep like qualities when it comes to blindly following what a newspaper says!

  • Anonymous

    The deaths were no doubt ‘attributed’ to the use of cannabis, but they more than likely died of unrelated causes. i.e., someone dies in a car crash, the autopsy reveals cannabis is present in the deceased individual’s system, and they conclude there is a connection.That is my speculation, anyways.

  • daniel campos

    Alcohol figures are wrong! They are counted by the thousands, not the hundreds!!!

    7,341 deaths in 2008 due to alcohol:

    8,724 deaths in 2007:

    8,724 alcohol-related deaths in 2007:

  • daniel campos

    Not only that, but the author gives the wrong reference for alcohol; he writes that it is from

    But if you go there and download the spreadsheet, you’ll see how it doesn’t include alcohol – what’s going on there?

  • daniel campos

    And there’s even more!!!:

    Also, in the spreadsheet that he uses for the statistics ( on cannabis and MDMA, it clearly states that are ” Number of deaths where selected substances were mentioned on the death certificate” – that’s why there are 19 for cannabis.


    It is very well known by everybody that number of deaths where ALCOHOL was mentioned on the death certificates is several thousands, not hundreds, and therefore this graph is not only bad, it is *extremely* bad.

  • big boss

    what the ,, prohibition peddling ,,,, 19 deaths for cannabis ,, prove it! ,
    i should think it’s saved more!!!!

  • Subba Rao

    This is a great information, ofcourse cannabis deaths are debatable. There are many theories behind this.