Random Analytica

Random thoughts, charts, infographics & analysis. Not in that order

Month: August, 2014

Random Analytics: Ebola in the DRC (to 27 Aug 2014)

Over the past 24-hours the World Health Organisation has confirmed another outbreak of Ebola Virus Disease (EVD), this time in the Democratic Republic of Congo. Although the Ebola species hasn’t yet been confirmed Ian M. Mackay (who knows a thing or two about this subject) suspects it might be Sudan Ebolavirus (read his excellent post and view his infographic here). To give you an idea where Equateur province is in relation to the DRC I’ve put together an infographic but have failed to locate where Ikanamongo Village is (where the index case ate the bush animal, possibly a bat and became ill).

According to my historical reckoning this would be the eighth outbreak of EVD in the DRC, six of Ebola Zaire, one of Ebola Bundibugyo and the most recent which remains unconfirmed. It should be noted that my count includes the retrospective confirmation of Ebola Zaire from 1972. If it does turn out to be a variant of Ebola Sudan then this would be the ninth iteration of the disease with 3 confirmations in what is now South Sudan, four in Uganda, one accidental infection in the UK and potentially the most recent DRC outbreak.

1 - Ebola_DRC_140828

Via the World Health Organisation. Ebola virus disease – Democratic Republic of Congo Disease outbreak news 27 August 2014. Excerpt:

Epidemiology and surveillance

On 26 August 2014, the Ministry of Health, Democratic Republic of Congo (DRC) notified the World Health Organization (WHO) of an outbreak of Ebola virus disease (EVD) in Equateur Province.

The index case was a pregnant woman from Ikanamongo Village who butchered a bush animal that had been killed and given to her by her husband. She became ill with symptoms of EVD and reported to a private clinic in Isaka Village. On 11 August 2014, she died of a then-unidentified haemorrhagic fever. Local customs and rituals associated with death meant that several health-care workers were exposed and presented with similar symptoms in the following week.

Between 28 July and 18 August 2014, a total of 24 suspected cases of haemorrhagic fever, including 13 deaths, have been identified. Human-to-human transmission has been established and includes the health-care personnel who were exposed to the deceased pregnant woman during surgery (one doctor and two nurses) in addition to the hygienist and a ward boy, all of whom developed symptoms and died. Other deaths have been recorded among the relatives who attended the index case, individuals who were in contact with the clinic staff, and those who handled the bodies of the deceased during funerals. The other 11 cases are currently being treated in isolation centres.

Samples have been sent to laboratories in Kinshasa and in Gabon for confirmation of EVD and to identify the strain. The index case and the 80 contacts have no history of travel to the EVD-affected countries in West Africa (Guinea, Liberia, Nigeria, or Sierra Leone) or history of contact with individuals from the affected areas. At this time, it is believed that the outbreak in DRC is unrelated to the ongoing outbreak in west Africa.


Data Sources

[1] d-maps.com. Africa: states, main cities. Accessed 28 August 2014.
[2] d-maps.com. Democratic Republic of the Congo / République Democratique du Congo: boundaries, provinces, main cities. Accessed 26 August 2014.
[3] Wikipedia. File:Flag of the Democratic Republic of the Congo.svg. Accessed 28 August 2014.
[4] World Health Organisation. Ebola virus disease – Democratic Republic of Congo Disease outbreak news 27 August 2014. Accessed 28 August 2014.

Random Analytics: Ebola in Liberia (to 19 Aug 2014)

It is noteworthy that over the past 24-hours we have received confirmation from the Liberian Ministry of Health and Social Welfare that the confirmed, probable and suspected cases of Ebola have now topped 1,000.

1 - Ebola_Liberia_140822

The Ebola in Liberia by County infographic details the cases and fatalities from Ebola. Each Country which is impacted by Ebola is listed with individuals listed as confirmed, probable, suspected then deceased. The provisional CFR is based on those total numbers.

Three Counties now have case counts in the hundreds including Lofa (455-cases), Monterrado (272-cases) which also includes the capital Monrovia and Bong (103-cases). Two Counties now have fatalities in the hundreds including Lofa (235) and Montserrado (212-cases).

The outbreak is ongoing so these numbers are still very provisional.

The other interesting find from doing this infographic was the Ministry of Health and Social Welfare data has some issues.

2 - Liberia_StatsByCounty_140822

I’ve highlighted the data anomalies (where confirmed, probable and suspected cases are less than the deaths). This could be an issue of how they have structured the original document (where case lines are opposite to death lines making it easy to input the wrong data).

I suspect that the analysts in the Ministry are under a lot of pressure. I’ll pass on my findings to ReliefWeb and the Liberian Ministry of Health and Social Welfare for their review and consideration.


Data Sources

[1] d-maps.com. Liberia / Republic of Liberia. Accessed 22 August 2014.
[2] Liberian Ministry of Health and Social Welfare. Liberia Ebola Sitrep no. 96. ReliefWeb. Accessed 22 August 2014.

Random Analytics: Ebola in Nigeria (to 16 Aug 2014)

ReliefWeb is reporting more cases in Liberia over the past 24-hours with data updated to 16 August 2014. The NIGERIA Daily Situation Report (SitRep No: 17) Date: 16th August 2014 includes a new confirmed case and three suspected cases. Crawford Kilian via his excellent H5N1 site has been following reports of a suspected case from Kaduna State, although this is not reported in the latest update (noting that said update is now almost three days old). The article coming out of Nigeria in relation to Kaduna can be found here and here while the counter-article can be found here.

As Crawford correctly points out Kaduna is an 80-minute flight away from Lagos where the imported outbreak is taking place. According to Google if you wanted to drive the 775-km it would take you about 10-hours.

01 - Ebola_Nigeria_140819

The above infographic details the cases and fatalities from Ebola in Nigeria. Cases include all suspected, probable and confirmed cases and the provisional CFR is based on those total numbers (whereas ReliefWeb uses only confirmed numbers for their CFR).

The only confirmed Nigerian State to be impacted by Ebola is Lagos from an imported case back in late July where there have been 12-confirmed cases (4 deceased with some now released and some still in quarantine). The other point I wanted to make out was that the most recent articles from Nigeria have mentioned both a case from Kaduna and effectively a retraction that could be the case.

We are still in the ‘Fog of the Outbreak’ thus I have highlighted Kaduna State and await more information and evidence.


Data Sources

[1] d-maps.com. Nigeria / Federal Republic of Nigeria, boundaries, states. Accessed 19 August 2014.
[2] ReliefWeb. NIGERIA Daily Situation Report (SitRep No: 17) Date: 16th August 2014. Accessed 19 August 2014.

Random Analytics: Ebola in Sierra Leone (to 14 Aug 2014)

Sherlock Holmes was famously quoted as saying that ‘when you have eliminated the impossible, whatever remains, however improbable, must be the truth?’

Yesterday I had a Twitter conversation with virologist Ian M Mackay and currently Sierra Leone deployed health reporter Jennifer Yang about the disconnect between the official World Health Organisation figures and the Sierra Leone Ministry of Health updates. In brief, the Sierra Leone MoH figures are always lower than those presented by WHO which made no sense to me given that the MoH should be data-prime. The numbers anomaly had recently been noted by key Flublogist Crawford Kilian on his blog H5N1.

During the conversation a number of theories were put forward by various parties including a lack of capacity on the ground, data-chain promulgation/speed issues and even a conspiracy theory where-by the Ebola outbreak is linked to central government suppression of the Kailahun/Kenema districts (which border both Guinea/Liberia) as both districts are opposition strongholds.

Although there is certainly issues around ‘boots on the ground’ capacity (see Jennifer’s excellent article from Sierra Leone) the answer to the question as to why there was a significant difference between the Sierra Leone MoH and the WHO numbers turned out to be blindingly obvious.

Most of the reports coming out directly from Sierra Leone are the Ministerial statements which only include confirmed numbers whereas if you dig a little further you can find the full updates including suspected and probable cases in the Ebola Situation Reports.

Here is the look at the situation in Sierra Leone by District based on the most updated Ebola Situation Report (Vol. 78 dated 14 August 2014).

01 - Ebola_SierraLeone_140815

The above infographic details the cases and fatalities from Ebola in Sierra Leone. Cases include all suspected, probably and confirmed cases and the provisional CFR is based on those total numbers.

As you can see the Kailahun and Kenema Districts are the most impacted regions with Sierra Leone accounting for 84.7% of all cases and 95.8% of all fatalities.


Data Sources

[1] Government of Sierra Leone Ministry of Health and Sanitation. EBOLA VIRUS DISEASE – SITUATION REPORT (Sit-Rep) – 14August, 2014. Government of Sierra Leone. Accessed 15 August 2014.


The Worsening Fatality Statistics in Australian Mining

For those that closely follow the Australian Mining Sector it will come as no surprise that 2014 is emerging as one of the worst in terms of safety that we have seen in a generation. According to SafeWork Australia in the first six months of this year there were 11 notifiable fatalities in the mining sector, which according to my calculations currently equates to a Worker Fatality Rate (WFR) of 28.8 fatalities per 100,000 workers. To give that some historical context the WFR for all Australian workers in 2013 was 1.64 fatalities per 100,000 workers.

2 - MiningFatalities_2003~2014

The first chart details the amount of work related fatalities by year since 2003. Figures exclude death by iatrogenic injuries, natural causes not related to work, disease, injuries sustained while overseas or suicide. The 2014 numbers are correct to 8 August with the 2013 and 2014 numbers reflecting the more comprehensive Industry of Workplace statistics (with thanks to the statistics team at SafeWork Australia for clarifying the differences).

Since the start of 2014 I have started to closely track employment, automation and fatalities as the three key indicators on the health of the mining sector. Within a few weeks I knew that mining safety would be a big story as a number of single fatalities occurred during January and February followed up by an underground collapse in April which killed two miners at the Austar Coal Mine .

Yet, a heightened fatality count in the mining industry isn’t the only story here.

Initially out of ignorance to how the industry and SafeWork Australia tracks its work related fatalities I started to build up a personal database of mining fatalities which also included those who have died of natural causes (on-site but not work related), from suicide, fatalities in overseas Australian miners and more recently those who could be considered Lifestyle Miners.

1 - MiningRelatedFatalities_2014

The second chart looks at Mining and Mining Related deaths of Australians in 2014.The WFR is calculated only on the official SafeWork Australia figures (correct as at 8 Aug 2014).

The key data point in the chart is the inclusion of known suicides. In June The West Australian mining industry was left reeling when an onsite incident led to the death of one employee and the possible offsite death of another. This incident has been followed up by two more probable onsite suicides amongst Pilbara FIFO workers. The recent tragedies come about as the District Coroner for the Pilbara region referred a number of 2013 deaths by suicide amongst FIFO workers to the WA State Coroner for a possible inquest.

It’s not all bad news though and I was heartened by news that AngloGold Ashanti, who were at the centre of the recent multiple tragedy in the Pilbara have in the past week signed up to the FIFO Families Social Support and Education Program.

In summary, I believe the mining industry must face the issue of mental health and suicide head-on. As far as I am concerned, if a miner dies by his or her own hand onsite should be treated in exactly the same way as if it were a work related fatality including the provision of industry wide data and statistics on the subject, more education to employees and their families and if required, seeking help from appropriate resources and organisations.


If you or someone you know is thinking of suicide, phone Lifeline on 13 11 14. Help is also available via Rural Link (1800 552 002), the Suicide Call Back Line (1300 659 467) and online resources can be found at BeyondBlue.

This article was originally published on MiningIQ.
Read the original article.

Data Sources

[1] Australian Bureau of Statistics. 6291.0.55.003 – Labour Force, Australia, Detailed, Quarterly, May 2014. Accessed 11 Aug 2014.
[2] SafeWork Australia. Worker fatalities. Accessed 11 Aug 2014.
[3] SafeWork Australia. Work-related Traumatic Injury Fatalities, Australia 2013. SafeWork Australia. 2014. Pg: iii-vii.

Random Analytics: The West African Ebola Outbreak (to 4 Aug 2014)

Here are some updated charts and infographics of the 2014 Ebola Virus Disease outbreak using a number of primary sources including the most recent World Health Organisation Disease Outbreak News (DON) released 6 August 2014.

***** Please note that all EVD infographics in this series were updated with public source information to 4 August 2014 *****

01 - Ebola_Top10OutbreaksByCaseNos_140808

Ebola (Top 10 Outbreaks by Case Numbers)

The first chart displays the top 10 outbreaks in order of case numbers. Each horizontal bar is filled with the flag(s) of the country where the outbreak occurred.

With clinical cases reaching 691 in Guinea, 516 in Liberia, 495 in Liberia and nine in Nigeria the West African outbreak has now become largest Ebola outbreak in history based on both case numbers (1711) and fatalities (932). The second largest outbreak was of the Ebola Sudan strain which occurred in Uganda (2000) when 425 became infected and 224 died. The recent outbreak is the first to migrate across international land borders. The only other recording of an EVD that jumped borders prior to this outbreak was the Gabon/RSA (1996) outbreak. In that instance a doctor caught the disease in Gabon and subsequently took an international flight to South Africa where he became ill and infected other Health Care Workers (HCWs).

02 - Ebola_CasesbyClassYear_140808

Ebola (Cases by Classification and Year)

The second chart shows cases by classification (in order they are Ebola Zaire, Sudan, Bundibugyo, Reston and Ivory Coast) by year and then split into those recovered or those deceased (which follows in a red variant). The West African outbreak has become the most significant in terms of case numbers, eclipsing the 1976 dual outbreaks which saw 603 cases and 431 deaths (a combined Case Fatality Rate of 71.5%).

Currently the provisional Western African outbreak has seen 1711 cases and 932 deaths (a CFR of 54.5%).

Notes: Several years had just one case. They are 1972 (a retrospective fatality of Ebola Zaire in Zaire), 1977 (a single case of Ebola Zaire in Zaire), 1988 (an accidental infection of Ebola Zaire in Porton Down, UK) and 2011 (a single fatality of Ebola Sudan in Uganda).

 03 - WestAfrica_Cases~FatalitiesMonth_140808

The West African Outbreak – Cases & Fatalities by Month

The final chart shows both case numbers and fatalities by month. Each column is split into the current four impacted countries with data represented by the varying national flags.

The very interesting data point that springs out from this chart is that the DON I utilised for this only had data for the first four days of the month yet cases are already 271 and fatalities are 106. It should be noted that those figures are not exact as the DON that covered the month rollover between July and August had to be estimated (using a 50/50% split).


Acknowledgements: Data for this infographic was sourced from official reports from the World Health Organisation. I have also utilised resources from the CDC, CIDRAP, FluTrackers, and H5N1. I’m also a big fan of the analytical work of Virology Down Under (Ian Mackay) and Mens et Manus (Maia Majumder).

Random Analytics: Australian Mining Employment (to July 2014)

Mining continues to play an important part in the overall economy of Australia. For all of the discussion about the sector many people don’t realise that mining only employs a fraction of Australia’s workforce (currently just 264,400 source: ABS) and that many in the industry work on the construction rather than the operational side.

Each month I spend some time collating stories from a wide range of industry and media sources to build some analytics around the current state of mining employment in Australia. Here are the charts for Australian Mining Employment through to the end of July 2014.

1 - MiningJobsByState_Infographic_Jun2014_140801

The opening infographic looks at job gains and losses by State or Territory for the month of July.

Data-points: After a horror month in June when 1,592 jobs were lost Western Australia has picked up some employment this month with announcements by Atlas Iron (+200) and Transalta (LNG +250). For the rest of the country the numbers were all in the negative, even in New South Wales which gained 450 (Whitehaven Coal) but lost 915 overall including big numbers in the Hunter Valley. Overall the nation gained 920 and lost 2,043 mining jobs.

2 - MiningGainsLosses_Chart_140801

The second employment charts looks at the previous 24-months from a total mining employment gain and loss perspective. The positive employment numbers are split into those that reflect infrastructure (tan) and operational (blue) gains. Job losses are represented in red.

The biggest data-point for July 2014 was that the big numbers seen last month reduced slightly as those some of those jobs finalised in July and that we are now in the third month of 1,000+ job losses and the second consecutive month of 2,000+ job losses. The last day of the month was especially bad with 400 Hastings Deering jobs going in Queensland and another 95 coal positions being axed by BHP in NSW.

It wasn’t all bad news as some large operational announcements, like the Whitehaven Coal announcement of 450 jobs at its Maules Creek mine offset some of the big losses being announced in the Hunter Valley.

3 - MiningResourceGainsLosses_Chart_140801

The next employment chart looks in more detail at the main resource types (Iron Ore, Coal, Gold/Copper, Zinc/Lead/Nickel, CSG/LNG and Uranium) by either a job gain or a job loss.

Two key data points:

  • Although there were some gains in both coal and iron ore they were much smaller than the operational jobs lost in both resource types;
  • Coal has lost operational jobs now for 18 consecutive months (the last recording of no job losses was in Jan 2013).

4 - MiningSectorSentiment_Chart_140801

The last chart tracks employment gains and losses sentiment and is now updated back to December 2011.

Although the employment news is quite bad the sentiment is actually positive. I’ve seen this occur before. If you look at the difference between Jul 2013 (-24) and August 2013 (+3) you see a big jump in sentiment. This occurred during last year’s commodity crash as government pushed through mining approvals (which is a positive indicator even though it might not include any immediate employment outcomes). So although there are continued job cut announcements they are being largely offset (in terms of sentiment) by positive future announcements, such as the Adani announcement for the build of Carmichael Coal in Queensland.

Juking the Stats (July 2014)

Silver Lake Resources is set to cut more jobs as the gold miner closes its Lakewood Mill but didn’t announce how many would go or when.


It’s been another bad month for mining employment which came as a surprise as often the first month in a Financial Year is a chance for companies and governments to highlight some good news or to make positive announcements. This is reflected in the data where job losses outweigh job gains but sentiment is improving.

Saying that governments don’t create jobs, businesses do and with so much infrastructure finalising and the narrative squarely fixed on productivity I’ll suggest that you will see more negative than positive news in coming months.