Decease data analysis
The detailed anaysis of the deceased is done based on the days to decease, days to report, and the place of death.
Histogram of Days to Decease(KARNATAKA)
The graph below represents the histogram of days to decease in Karnataka based on the data collected, we separated only the patients who were diagnosed in hospitals and then the difference of admission date and decease date was found for each patient.Then we plot a frequency histogram.
More than 4000 people die just after first day of hospitalisation and most of the deceased patients who occupied hospital beds were dead within day 2 weeks,and and the mean line is shown, which is nearly about 6 days.This tells us that fatalities were mostly within 2 weeks after the patient was hospitalised.
Scatter plot for Days to Decease(KARNATAKA)
While plotting this, we took all districts except Bengaluru Urban.Since Bengaluru Urban had high number of death counts, so to have a better understanding among the other districts we left out Bengaluru urban and carried our analysis.
It is evident from the plot that in almost all the districts, fatalities were within 2 weeks.
Scatter plot for Days to Decease(KARNATAKA)
Scatter plot for all the districts are done here in the following plot.
It is evident that Bengaluru Urban is unique because it had high number of death counts.Here to the point are dense very early in the x-axis implying that fatalities were mostly within 2-weeks.
Histogram of Days to Report(KARNATAKA)
The graph below represents the distribution of days to report in Karnataka based on the data collected, from the Media Bulletin published by the state government, we calculated the difference of decease date and media bulletin date for each patient and we call the difference as days to report and subsequently plot the frequency histogram.
For most of deceased patients, reporting is quick as the histogram slides down very quickly and also the mean line is shown which is nearly 7 days
Scatter plot for Days to Report(KARNATAKA)
While plotting this, we took all districts except Bengaluru Urban.Since Bengaluru Urban had high number of death counts, so to have a better understanding among the other districts we left out Bengaluru urban and carried our analysis.
From the above plot we observe that in the points are very dense within day 7,i.e,most of the districts reported death within a week.
Scatter plot for Days to Report(KARNATAKA)
Scatter plot for all the districts are done.
Bengaluru Urban is different from the rest as it had high death counts.
Barplot of Place of Death
The graph below represents the Barplot of Place of Death in Karnataka based on the data collected.As per the Media Bulletin published by the Karnataka state government, the individual deaths were classified into three categories,viz.,Hospitalised,Brought Dead,and Died at residence,we took the category counts and plot a barplot.
Most of the patients were actually diagnosed in hospitals before they were fatal to covid-19.
Barplot of Place of Death(District wise)
The graph below represents the Barplot of Place of Death(District wise) in Karnataka based on the data collected.
Most of the patients were actually diagnosed in hospitals before death.But Bengaluru urban shows a peak in death count in hospitals.
Analysing Wave-Wise
The entire data collected has been analysed wave wise to get a better understanding of the impact of Covid-19 and the nature of victims.
- Wave 1 - till the end of October 2020
- Wave Middle -From Nov-2020 to Jan-2021
- Wave 2 - since the end of January 2021 till present
Since,the counts during the Wave Middle were comparatively low,so,we only considered Wave 1 and Wave 2 for our analysis..
Histogram of Days to Decease(Wave-1 vs Wave-2)
The graph below represents the distribution of days to decease(wave-wise) in Karnataka based on the data collected.
The trend in both the wave is almost similar but since Wave 2 had more death counts, it is also evident here in the frequencies.Also the mean line is shown for both the waves, one can also perform a t-test to validate that the means are not equal for the waves.
Histogram of Days to Report(Wave-1 vs Wave-2)
The graph below represents the distribution of days to report(wave-wise) in Karnataka based on the data collected.
Both wave shows similar trend but the counts in wave 2 is high because of high number of death counts.Also the mean line is shown for both the waves, one can also perform a t-test to validate that the means are not equal for the waves.
Confidence Interval for Days to Decease(Wave-1 vs Wave-2)
The graph below represents the confidence interval for days to decease(wave-wise) in Karnataka based on the data collected.First,the NA's,and Others in the district column were omitted, then 95% t-interval around the mean is calculated and plotted below.
Mean Scatter plot for Days to Decease(Wave-1 vs Wave-2)
The graph below represents the Mean scatter plot for days to decease(wave-wise) in Karnataka based on the data collected.First,the NA's,and Others in the district column were omitted, for each individual district, we calculated the mean days to decease wave-wise and then a scatter plot is plotted along with the line y=x.The blob size indicates the total number of cases in that district for both waves.
.
Confidence Interval for Days to Report(Wave-1 vs Wave-2)
The graph below represents the confidence interval for days to report(wave-wise) in Karnataka based on the data collected.First,the NA's,and Others in the district column were omitted, then 95% t-interval around the mean is calculated and plotted below.
.
Mean Scatter plot for Days to Report(Wave-1 vs Wave-2)
The graph below represents the Mean scatter plot for days to Report(wave-wise) in Karnataka based on the data collected.First,the NA's,and Others in the district column were omitted, for each individual district, we calculated the mean days to report wave-wise and then a scatter plot is plotted along with the line y=x.The blob size indicates the total number of cases in that district for both waves.
.
Box-Plot for Days to Decease
The graph below represents the Box-plot for days to decease in Karnataka(District-wise) based on the data collected.All District shows similar trend.
Box-Plot for Days to Report
The graph below represents the Box-plot for days to report in Karnataka(District-wise) based on the data collected.Most districts have similar trend but some have high interquartile range along with high number of outliers.
Box-Plot for Days to Decease Age-wise
Box-Plot for Days to Decease Age-Gender-wise
Scatter plot for Days to Decease Age-wise
Scatter plot for Days to Decease Age-Gender-wise
Confidence Interval for Days to Decease Age-wise
Confidence Interval for Days to Decease Age-Gender-wise
From the media bulletin, we took the difference between the date of decease and admission date and call the difference as days to decease.We then separated the deceased on the basis of age and gender and calculated the mean, variance, and standard deviation for days to decease.We then plot the 95% t-confidence interval around the mean for all the age-gender category.
Confidence Interval for Days to Report Age-Gender-wise
From the media bulletin published by the state government, we took the difference between the date of decease and the media bulletin date and call the difference as the days to report.We then separated the deceased on the basis of age and gender and calculated the mean, variance, and standard deviation for days to report.We then plot the 95% t-confidence interval around the mean for all the age-gender category.
Barplot of Place of Death(Wave-1 vs Wave-2)
The graph below represents the Barplot of Place of Death(wave-wise) in Karnataka based on the data collected.We first separated each individual deceased patient with respect to wave and then then took counts on the basis of place of death.
The counts in wave-2 are higher because of the higher number of fatalities.
Scatter plot of District-wise Hospitalisation Rate.
The graph below represents the scatter plot of districts in Karnataka based on the data collected.Here we take the ratio of patients hospitalised(i.e,we take the counts of patients who were hospitalised and divide it by total number of death across districts) on one axis and plot a scatter plot with total deaths on the other.This gives us a understanding of how well hospitalisation is done in a particular district.
Most of the districts did well in hospitalising patients.Since most of the districts are above the .95 mark.
Scatter plot of Hospitalisation Rate(Wave 1 vs Wave 2).
The scatter plot of Hospitalisation Ratio of the deceased patients has been presented here across districts.The Media Bulletin reports each deceased patient as either admitted to hospital, or brought dead or died at residences, we call both the categories brought dead and died at residence as deceased patients who were not hospitalised.Then for each district, we count the total deceased patients and total hospitalised patients and take the ratio of hospitalised patients to total deceased patients to arrive at the hospitalisation ratio.We calculate the ratio for each district wave wise and then plot plot a scatter plot along with line y=x.
Scatter plot of District-wise Hospitalisation Rate(Subplots).
Now for a better analysis we clustered individual districts on the basis of total death counts and then we again perform the similar analysis as earlier.Here we took 3 categories.Firstly with total death count less than 500,secondly total death count ranging within 500 to 1500,and thirdly districts with death count higher than 1500.
