Message Boards

GROUPS:

Staff Picks Biological Sciences Data Science Medical Sciences Graphics and Visualization Wolfram Language Events & Media

Oscar Rodriguez

General and COVID-19 deaths in Sweden

Oscar Rodriguez, corporación universitaria iberoamericana

Posted 9 months ago

2881 Views

7 Replies

12 Total Likes

Follow this post

MODERATOR NOTE: coronavirus resources & updates: https://wolfr.am/coronavirus

General and COVID-19 deaths in Sweden
Oscar Rodriguez, Corporación Universitaria Iberoamericana

In this notebook i am trying to uncover if the sar-cov-2 is having a major impact in deceases in Swedish. I am using the same database previously used by Jan Brugard (https://community.wolfram.com/groups/-/m/t/1974412).

Death by day

First, i would like to see how the deceased are distributed since 2015. I am excluding the day with zero deaths

deathByDay=Import[FileNameJoin[{NotebookDirectory[],"2020-05-08---preliminar-statistik-over-doda-inkl-eng.xlsx"}],{"XLSX","Data",2}];(*NumberofdeathsreportedtoStatisticsSweden,perdayfor2015-20201(entirecountry)*)

TableForm[deathByDay,TableHeadingsAutomatic];

of2015=DeleteCases[deathByDay[[8;;,2]],0.`];(*seborranlosregistrosconcero,yaqueelinteressonelconteodelosfallecidos*)

of2016=DeleteCases[deathByDay[[8;;,3]],0.`];

of2017=DeleteCases[deathByDay[[8;;,4]],0.`];

of2018=DeleteCases[deathByDay[[8;;,5]],0.`];

of2019=DeleteCases[deathByDay[[8;;,6]],0.`];

of2020=DeleteCases[deathByDay[[8;;,7]],0.`];

ofAverage=deathByDay[[8;;,8]];

DistributionChart[{of2015,of2016,of2017,of2018,of2019,of2020,ofAverage},ChartElementFunctionChartElementData["PointDensity"],ChartLabels{"2015","2016","2017","2018","2019","2020","Average"},PlotLabel"Number of deaths reported to Statistics Sweden,per day for 2015-20201 (entire country)",ChartStyle"SouthwestColors",BaseStyle14,ImageSize800,ScalingFunctions"Log",FrameLabel{"Year","# of death"}]

This graph shows the distribution of the death count from 2015 to 2020. In all cases, there is a very marked distribution of deaths, with some peaks in all years except 2020. This is because all the data is not available. in 2020, only until May 6. In 2020 there is a more dispersed trend.

Este gráfico muestra la distribución del conteo de muertes desde al año 2015 al 2020. En todos los casos se ve una distribución de las muertes muy marcado, con algunos picos en todos los años excepto en 2020. Esto debido a que no están todos los datos en 2020, solo hasta mayo 6. En 2020 se ve una tendencia más dispersa.

DistributionChart[{of2019,of2020},ChartLayout"Overlapped",ChartBaseStyleOpacity[0.3],ChartLegends{"otro","2020"},PlotLabel"Number of deaths reported to Statistics Sweden,per day for 2015-20201 (entire country)",ChartStyle"SouthwestColors",BaseStyle14,ImageSize800,FrameLabel{"Year","# of death"}]

Now, this is showing the whole year, except for 2020. So this is how it looks if we show the deceased until may 6 every year.

LocationEquivalenceTest[{Standardize[of2015May6],Standardize[of2020May6]}]

0.469462

of2015May6=DeleteCases[deathByDay[[8;;134,2]],0.`];(*seborranlosregistrosconcero,yaqueelinteressonelconteodelosfallecidos*)

of2016May6=DeleteCases[deathByDay[[8;;134,3]],0.`];

of2017May6=DeleteCases[deathByDay[[8;;134,4]],0.`];

of2018May6=DeleteCases[deathByDay[[8;;134,5]],0.`];

of2019May6=DeleteCases[deathByDay[[8;;134,6]],0.`];

of2020May6=DeleteCases[deathByDay[[8;;134,7]],0.`];

ofAverageMay6=deathByDay[[8;;134,8]];

DistributionChart[{of2015May6,of2016May6,of2017May6,of2018May6,of2019May6,of2020May6,ofAverageMay6},ChartElementFunctionChartElementData["PointDensity"],ChartLabels{"2015","2016","2017","2018","2019","2020","Average"},PlotLabel"Number of deaths reported to Statistics Sweden,per day for 2015-20201 (entire country) until may 6",ChartStyle"SouthwestColors",BaseStyle14,ScalingFunctions"Log",ImageSize900,FrameLabel{"Year","# of death"}]

This graph shows the distribution of deaths until May 6, in this case the peaks of deaths are not evident in all the years prior to 2020.

Esta gráfica muestra la distribución de las muertes hasta mayo 6, en este caso no se evidencian los picos de muertes en todos los años previos al 2020.

probabli=SmoothHistogram[{of2015May6,of2016May6,of2017May6,of2018May6,of2019May6,of2020May6},PlotRangeAll,PlotLegends{"2015","2016","2017","2018","2019","2020","Average"},PlotLabel"Death distribution 2015-2020 until may 6",PlotTheme"Web"];

LocationEquivalenceTest[{of2015May6,of2016May6,of2017May6,of2018May6,of2019May6,of2020May6},{{"TestDataTable",All},"TestConclusion"}]

LocationEquivalenceTest

:No valid tests were found under the given assumptions. The KruskalWallis test will be used but may produce unreliable results.



	Statistic	P-Value
Kruskal-Wallis	93.3267	8.0296× -20 10

,The null hypothesis that the mean are equal is rejected at the 5 percent level based on the Kruskal-Wallis test.

cumula=SmoothHistogram[{of2015May6,of2016May6,of2017May6,of2018May6,of2019May6,of2020May6},Automatic,"CDF",PlotRangeAll,PlotLegends{"2015","2016","2017","2018","2019","2020"},PlotLabel"Death distribution 2015-2020 until may 6, cumulative",PlotTheme"Web"];

survi=SmoothHistogram[{of2015May6,of2016May6,of2017May6,of2018May6,of2019May6,of2020May6},Automatic,"SF",PlotRangeAll,PlotLegends{"2015","2016","2017","2018","2019","2020"},PlotLabel"Death distribution 2015-2020 until may 6, survival",PlotTheme"Web"];

Grid[{{probabli},{cumula},{survi}},ItemSize50]

	2015
	2016
	2017
	2018
	2019
	2020

	2015
	2016
	2017
	2018
	2019
	2020

	2015
	2016
	2017
	2018
	2019
	2020

The distributions between the different years do not differ much. However, when generating an inferential statistical test, it was not possible to find a valid test.

Las distribuciones entre los diferentes años no difieren mucho. Sin embargo, al generar una prueba estadística inferencial, no fue posible encontrar un test válido.

I am going to try to use inferential statistics, in order to determine if the differences are significant

deathBySex=Import[FileNameJoin[{NotebookDirectory[],"2020-05-08---preliminar-statistik-over-doda-inkl-eng.xlsx"}],{"XLSX","Data",3}];

TableForm[deathBySex,TableHeadingsAutomatic];

(*NumberofdeathsreportedtoStatisticsSweden,perdayfor2015-20201(entirecountry)*)

(*mensegmentedbyage*)

men0to64=DeleteCases[deathBySex[[9;;,3]],0.`];

men65to79=DeleteCases[deathBySex[[9;;,4]],0.`];

men80to89=DeleteCases[deathBySex[[9;;,5]],0.`];

men90to=DeleteCases[deathBySex[[9;;,6]],0.`];

(*womensegmentedbyage*)

wom0to64=DeleteCases[deathBySex[[9;;,7]],0.`];

wom65to79=DeleteCases[deathBySex[[9;;,8]],0.`];

wom80to89=DeleteCases[deathBySex[[9;;,9]],0.`];

wom90to=DeleteCases[deathBySex[[9;;,10]],0.`];

DistributionChart[{Labeled[{men0to64,wom0to64},"No test",Above],Labeled[{men65to79,wom65to79},LocationEquivalenceTest[{men65to79,wom65to79}],Above],Labeled[{men80to89,wom80to89},LocationEquivalenceTest[{men80to89,wom80to89}],Above],Labeled[{men90to,wom90to},"No test",Above]},ChartLabels{Placed[{"0-64","65-79","80-89","90+"},Axis],Placed[{"","","",""},Axis]},ChartElementFunctionChartElementData["PointDensity"],PlotLabel"Number of deaths reported to Statistics Sweden, men and women, 2019 (entire country)",FrameLabel{"Age","# of death"},ChartLegends{"Men","Women"},BaseStyle14,ImageSize800,ScalingFunctions"Log",ChartStyle"SouthwestColors"]

LocationEquivalenceTest[{men0to64,wom0to64},{{"TestDataTable",All},"TestConclusion"}]

LocationEquivalenceTest

:No valid tests were found under the given assumptions. The KruskalWallis test will be used but may produce unreliable results.



	Statistic	P-Value
Kruskal-Wallis	219.561	1.28805× -58 10

,The null hypothesis that the mean difference is 0 is rejected at the 5 percent level based on the Kruskal-Wallis test.

This figure shows a comparison between men and women deceased in 2019. The number in the upper part is the p-value of the comparison, “No test” means that the LocationEquivalenceTest did not find a valid test to compare the two vectors, like shown up here. There are significant variations in the age rank: 65-79 and 80-89

Esta figura muestra una comparación entre hombres y mujeres fallecidos en 2019. El número en la parte superior es el valor p de la comparación, “Sin prueba” significa que la Prueba de equivalencia de ubicación no encontró una prueba válida para comparar los dos vectores, como se muestra aquí Hay variaciones significativas en el rango de edad: 65-79 y 80-89

men0to64in2020=DeleteCases[deathBySex[[9;;,12]],0.`];

men65to79in2020=DeleteCases[deathBySex[[9;;,13]],0.`];

men80to89in2020=DeleteCases[deathBySex[[9;;,14]],0.`];

men90toin2020=DeleteCases[deathBySex[[9;;,15]],0.`];

wom0to64in2020=DeleteCases[deathBySex[[9;;,16]],0.`];

wom65to79in2020=DeleteCases[deathBySex[[9;;,17]],0.`];

wom80to89in2020=DeleteCases[deathBySex[[9;;,18]],0.`];

wom90toin2020=DeleteCases[deathBySex[[9;;,19]],0.`];

DistributionChart[{Labeled[{men0to64in2020,wom0to64in2020},"No test",Above],Labeled[{men65to79in2020,wom65to79in2020},LocationEquivalenceTest[{men65to79in2020,wom65to79in2020}],Above],Labeled[{men80to89in2020,wom80to89in2020},LocationEquivalenceTest[{men80to89in2020,wom80to89in2020}],Above],Labeled[{men90toin2020,wom90toin2020},"No test",Above]},ChartLabels{Placed[{"0-64","65-79","80-89","90+"},Axis],Placed[{"","","",""},Axis]},ChartElementFunctionChartElementData["PointDensity"],PlotLabel"Number of deaths reported to Statistics Sweden, men and women, 2020 (entire country)",FrameLabel{"Age","# of death"},ChartLegends{"Men","Women"},BaseStyle14,ScalingFunctions"Log",ImageSize700,ChartStyle"SouthwestColors"]

This figure shows a comparison between men and women deceased in 2020 The number in the upper part is the p-value of the comparison, “No test” means that the LocationEquivalenceTest did not find a valid test to compare the two vectors. There are significant variations in the age rank: 65-79 and 80-89.

Esta figura muestra una comparación entre hombres y mujeres fallecidos en 2020 El número en la parte superior es el valor p de la comparación, “Sin prueba” significa que la Prueba de equivalencia de ubicación no encontró una prueba válida para comparar los dos vectores. Hay variaciones significativas en el rango de edad: 65-79 y 80-89.

Now it will be great to compare 2019 vs 2020

DistributionChart[{Labeled[{men0to64,men0to64in2020},"No test",Above],Labeled[{men65to79,men65to79in2020},"No test",Above],Labeled[{men80to89,men80to89in2020},"No test",Above],Labeled[{men90to,men90toin2020},"No test",Above]},ChartLabels{Placed[{"0-64","65-79","80-89","90+"},Axis],Placed[{"","","",""},Axis]},ChartElementFunctionChartElementData["PointDensity"],PlotLabel"Number of deaths reported to Statistics Sweden, men 2019 vs 2020 (entire country)",FrameLabel{"Age","# of death"},ChartLegends{"Men 2019","Men 2020"},ScalingFunctions"Log",BaseStyle14,ImageSize700,ChartStyle"SouthwestColors"]

LocationEquivalenceTest[{men90to,men90toin2020},{{"TestDataTable",All},"TestConclusion"}]

LocationEquivalenceTest

:No valid tests were found under the given assumptions. The KruskalWallis test will be used but may produce unreliable results.



	Statistic	P-Value
Kruskal-Wallis	51.2679	2.0419× -13 10

,The null hypothesis that the mean difference is 0 is rejected at the 5 percent level based on the Kruskal-Wallis test.

This figure shows a comparison between men 2019 vs 2020. The number in the upper part is the p-value of the comparison, “No test” means that the LocationEquivalenceTest did not find a valid test to compare the two vectors.

Esta figura muestra una comparación entre hombres 2019 vs 2020. El número en la parte superior es el valor p de la comparación, “Sin prueba” significa que la Prueba de Equivalencia de Ubicación no encontró una prueba válida para comparar los dos vectores.

DistributionChart[{Labeled[{wom0to64,wom0to64in2020},LocationEquivalenceTest[{wom0to64,wom0to64in2020}],Above],Labeled[{wom65to79,wom65to79in2020},"No test",Above],Labeled[{wom80to89,wom80to89in2020},"No test",Above],Labeled[{wom90to,wom90toin2020},"No test",Above]},ChartLabels{Placed[{"0-64","65-79","80-89","90+"},Axis],Placed[{"","","",""},Axis]},ChartElementFunctionChartElementData["PointDensity"],PlotLabel"Number of deaths reported to Statistics Sweden, men and women,2019 vs 2020 (entire country)",FrameLabel{"Age","# of death"},ChartLegends{"Women 2019","Women 2020"},ScalingFunctions"Log",BaseStyle14,ImageSize700,ChartStyle"SouthwestColors"]

LocationEquivalenceTest[{wom90to,wom90toin2020},{{"TestDataTable",All},"TestConclusion"}]

LocationEquivalenceTest

:No valid tests were found under the given assumptions. The KruskalWallis test will be used but may produce unreliable results.



	Statistic	P-Value
Kruskal-Wallis	62.5705	3.1749× -16 10

,The null hypothesis that the mean difference is 0 is rejected at the 5 percent level based on the Kruskal-Wallis test.

This figure shows a comparison between women 2019 vs 2020. The number at the top is the p-value of the comparison, “No test” means that the Location Equivalence Test did not find a valid test to compare the two vectors. Comparison is observed in the 0-64 age range without significance.

Esta figura muestra una comparación entre mujeres 2019 vs 2020. El número en la parte superior es el valor p de la comparación, “Sin prueba” significa que la Prueba de Equivalencia de Ubicación no encontró una prueba válida para comparar los dos vectores. Se observa comparación en el rango de edad 0-64 sin significancia.

By sex and age

data=Import[FileNameJoin[{NotebookDirectory[],"2020-05-08---preliminar-statistik-over-doda-inkl-eng.xlsx"}],{"XLSX","Data",10}];

TableForm[data,TableHeadingsAutomatic];

(*Deathofmenin2015segmentedbyage-Muertedehombresen2015segmentadosporedad*)

men0to64in2015=data[[2;;54,5]];

men65t079in2015=data[[55;;107,5]];

men80to89in2015=data[[108;;160,5]];

men90plusIn2015=data[[161;;213,5]];

(*Deathofmenin2015segmentedbyage,untilMay6-Muertedehombresen2015segmentadosporedad,hastamayo6*)

men0to64in2015may6=data[[2;;20,5]];

men65t079in2015may6=data[[55;;73,5]];

men80to89in2015may6=data[[108;;126,5]];

men90plusIn2015may6=data[[161;;179,5]];

(*Deathofwomenin2015segmentedbyage-Muertedemujeresen2015segmentadosporedad*)

wom0to64in2015=data[[214;;266,5]];

wom65to79in2015=data[[267;;319,5]];

wom80to89in2015=data[[320;;372,5]];

wom90plusIn2015=data[[373;;425,5]];

(*Deathofwomenin2015segmentedbyage,untilMay6-Muertedemujeresen2015segmentadosporedad,hastamayo6*)

wom0to64in2015may6=data[[214;;232,5]];

wom65to79in2015may6=data[[267;;285,5]];

wom80to89in2015may6=data[[320;;338,5]];

wom90plusIn2015may6=data[[373;;391,5]];

(*Deathofmenin2020segmentedbyage-Muertedehombresen2020segmentadosporedad*)

men0to64in2020=data[[426;;444,5]];

men65to79in2020=data[[445;;463,5]];

men80to89in2020=data[[464;;482,5]];

men90plusIn2020=data[[483;;501,5]];

(*Deathofwomenin2020segmentedbyage-Muertedemujeresen2020segmentadosporedad*)

wom0to64in2020=data[[502;;520,5]];

wom65to79in2020=data[[521;;539,5]];

wom80to89in2020=data[[540;;558,5]];

wom90plusIn2020=data[[559;;577,5]];

DistributionChart[{Labeled[{men0to64in2015,men0to64in2020},"No test",Above],Labeled[{men65t079in2015,men65to79in2020},"No test",Above],Labeled[{men80to89in2015,men80to89in2020},"No test",Above],Labeled[{men90plusIn2015,men90plusIn2020},"No test",Above]},ChartLabels{Placed[{"0-64","65-79","80-89","90+"},Axis],Placed[{"","","",""},Axis]},ChartElementFunctionChartElementData["PointDensity"],PlotLabel"Number of deaths reported to Statistics Sweden, per week, MEN 2015-2019 and 20201 (entire country)",ChartStyle"SouthwestColors",FrameLabel{"Age","# of death"},ChartLegends{"men 2015-2019","men 2020"},BaseStyle14,ImageSize700,ScalingFunctions"Log"]

This figure shows a comparison between men 2019 vs 2020 weekly. The number in the upper part is the p-value of the comparison, “No test” means that the LocationEquivalenceTest did not find a valid test to compare the two vectors.

LocationEquivalenceTest[{men0to64in2015,men0to64in2020},{{"TestDataTable",All},"TestConclusion"}]

LocationEquivalenceTest

:No valid tests were found under the given assumptions. The KruskalWallis test will be used but may produce unreliable results.



	Statistic	P-Value
Kruskal-Wallis	0.0980247	0.756656

,The null hypothesis that the mean difference is 0 is not rejected at the 5 percent level based on the Kruskal-Wallis test.

LocationEquivalenceTest[{men65t079in2015,men65to79in2020},{{"Te