A data based approach using suicide rates from 1985 to 2016
Introduction
The follwing story is based on a dataset about a suicide rates overview from 1985 to 2016. The shown story will only include data up to 2015 due to inconsistent data availablility in 2016. Nevertheless, it has some gainful insights for the readers that will catch the attention in due course.
The dataset covers more than 27'000 rows and in total 12 columns. It has data from all the countries worldwide and gives a global overview (with at maximum 90 countries per year) of suicide statistics. Furthermore, the dataset covers information about the GDP of countries and the Human Development Index (HDI). For this analysis, the HDI data has been ignored due to lot of missing data.
The story is structured into three parts and a conlcusion. The first part shows a rough overview of the global trend without diving into many details. The second paragraph backs the macro-level by looking into any trends on a global scale. The last part discusses the micro area. Different countries will be highlighted by displaying the lower and the upper tail countries based on suicide rates within the data.
Part I: How does the total amount of suicides compare with the ratio of total suicides/100'000 (100k) per population?
To get a first rough overview, I was summing up all data across the countries to get a total for each year. This first line graph gives the viewer a solid understanding about the indicating trend.
A line chart is one of the simplest visualization techniques but also a very efficient one. The chart above tells us the trend was increasing quite a lot from the mid 1980s with around 12 million deaths worldwide and plateaued at the beginning of 2000 with approximately 25 million suicides. After that, a small decrease can be observed. One catch the reader should also take into account is that not all countries were reporting figures each year for the whole time period and over the time more and more countries delivered reliable figures according to the dataset.
One obvious disadvantage of having total death’s counts is that it does not include the country population. The world has experienced a constant growth of its population worldwide especially within the timespan from 1985 to 2015. Thus, for the rest of the shown analysis, the ratio of total death’s count divided by 100k of the population will be considered.
With the help of the line chart above, the picture can be backed with more details. We can see the same trend — first increasing and after that a small decline — as in the first figure with the total suicide count. But the decreasing trend is easier visible. Additionally, it can be stated that at the maximum, a value of around 14'000 deaths could be counted per 100k in the population, while in the 1980s the ratio was constantly below 10%.
Part II: Is there any trend to see overall on a global scale considering macro-level playing fields?
This paragraph will show you details about specific trends in different generations and genders globally. The first part showed a constant increase of the trend untill the beginning of 2000 and after that a modest ongoing decline (see Figures 1 and 2).
Below, I analysed first the break down of the ratio of the suicides/100k per population by gender. This is one of my main points of the study:
Gender view
Men over-proportionally kill themselves more often than females according to the dataset. That overall outcome persists throughout the whole time period and is an interesting catch in my opinion which I did not had in mind before conducting the analysis.
The trend of female suicides seems to be constantly at about 2'000 per 100k per population globally without any much volatility.
On the other hand, the men’s count is at least by a factor of 2 and more. At the maximum, the factor was five-fold in 1995 compared to the females. Furthermore, the trend is more volatile and as a positive note: The trend is decreasing since the beginning of 2000s.
Age group view
A second analysis step included the different age groups of the study. This is also eye-catching in my view.
The bar chart in figure 4 shows many interesting findings. I will only emphasize two main insights. Firstly, the age range of 75+ years contributes the most to the suicides over the time span.
The second point is the steady consolidation of the suicides over time. The trend is narrowing down for all age groups. This can be seen as a positive sign.
Part III: Who are the most and least countries contributing to the suicide ratio over time?
My last point discusses the time series on the micro level of the countries. I was interested to see which countries especially could be found on the upper and lower tails of the suicide contribution. I defined a tail as ten countries. The time ranges can be summarized into 4 different static time ranges:
1985
1995
2005
2015
The lower tails over the time trend
The interpretation of the lower bars are partially insigthful. The countries with the smallest ratios seem to be islands, pretty small according to the land -or population size or are located in the middle hemispheres. If this can be taken as a reasonable finding — I would doubt it.
The upper tail over the time trend
In my opinion, the upper tail analysis gives more interesting view points. What can be seen in the bar charts above in figure 6 is a possible implication of the lifted iron curtail at the beginning of the 1990s. A lot of former Soviet Union countries e.g. Lithuania, Kazakhstan or Ukraine can be constantly found in the statistics.
Conclusion
In my first paragraph, you could see the overall trend of suicide rates was first increasing untill the 2000s and reversing afterwards which is a good sign.
The male contribution is over-proportionally higher than the female’s and this trend is consistent over time. Furthermore, the age range of +75 has the most impact on the suicide rate. Those were my second analysis outcomes.
In the last part, the lower tails give modest implications about the countries and its location, while the upper tails tell a story of possible iron curtail implications as quite a lot of cuntries of the former Soviet Union can be found on the list after it broke apart.
What has to be kept in mind of the shown analysis is that not all countries reported their numbers across the whole time range. There were at maximum around 90 countries in the dataset present and at minimum approximately 50 countries for each year of the study.
As a next action step for interested readers could be to consider the HDI data and the economic factors e.g. the GDP figures which are available in the dataset to conduct further analysis and gain more details.
Here you can find the link to my Github repo: https://github.com/Tobinho91/suicides_rates_1985_2015.git
