Clusters of Symptoms, Co-Morbidities and Description of Deceased across districts
In the media bulletin in the Symptoms column the symptoms observed in each deceased patient is listed. To see the variation in symptoms of deceased patients across different districts, we first grouped symptoms into 10 clusters. The specific clustering is provided in the following csv file.
This CSV file contains information about the different symptom clusters and the different sets of symptoms that are classified under each symptom cluster. In the CSV file, the 10 columns represent the 10 symptom clusters and the heading of each column is the name of that symptom cluster. The rows below the headings represent the different symptoms classified under the corresponding symptom clusters.
In the above graph we have plotted a heat map of the percentage(within each district) number of deceased patients across symptom clusters for all districts. Every tick on the x-axis represents the different symptom clusters. The y-axis represents the different districts. Each cell represents the percentage number of deceased patients in that symptom clusters for a district. The darker the colour of a cell, the smaller the percentage of deaths in that symptom for the given district while the brighter the colour, the larger is the corresponding percentage.
From the above symptom clusters heat maps, it can be inferred that in any given district of Karnataka, the percentage of deceased patients suffering from symptoms of breathlessness, fever, cold or cough is much higher than the percentage of patients suffering from any other symptoms. Thus, breathlessness, fever, cold and cough are the most important symptoms observed in deceased patients. Other than these four major symptoms, it can also be inferred that a significant number of deceased patients were either asymptomatic, had body pain and fatigue, chest pain or abdominal pain related symptoms. Very few patients suffered from a loss of consciousness or altered senses.
In the media bulletin for each deceased patient the source of illness is identified in the Description column. We have grouped the sources into 4 travel related clusters, SARI, ILI and Unknown. The specific clustering is provided in the following csv file.
Clusters of Description of Source of Illness
This CSV file contains information about the clusters of sources of illness based on the description associated with each patient in the Description column of the Media Bulletin. In the CSV file, the 7 columns represent the 7 clusters of description and the heading of each column is the name of that cluster of description of the source of illness. The rows below the headings are the description of the sources of illness classified under the corresponding cluster of description of the sources of illness.
To see the variation in description of source of illness we have plotted a heat map of the percentage (within each district) number of deceased patients across description clusters for all districts. Every tick on the x-axis represents the different source of illness. The y-axis represents the different districts. Each cell represents the percentage number of deceased patients in that description cluster for that particular district. The darker the colour of a cell, the smaller the percentage of deaths in that description for the given district while the brighter the colour, the larger is the corresponding percentage.
From the above heat maps it can be inferred that in any given district, the number of deceased patients who classified under ILI, SARI or whose investigation and tracing was still ongoing are much higher than the deceased patients classified under other descriptions. In each district, the percentage of patients classifying under the cluster Severe Acute Respiratory Illness is the highest making it the most important cluster of description.
In the media bulletin in the Comorbidities column the Comoorbidities observed for each deceased patient is listed. To see the variation in comorbidities of deceased patients across different districts, we first grouped the comorbidities into 15 clusters. The specific clustering is provided in the following csv file.
This CSV file contains information about the different clusters of comorbidities and the different sets of comorbidities that are classified under each comorbidity cluster. In the CSV file, the 15 columns represent the 15 comorbidities clusters and the heading of each column is the name of that comorbidity cluster. The rows below the headings are the comorbidities classified under the corresponding clusters of comorbidities
In the above graph we have plotted a heat map of the percentage(within each district) number of deceased patients across comorbidities clusters for all districts. Every tick on the x-axis represents the different comorbidities clusters. The y-axis represents the different districts. Each cell represents the percentage number of deceased patients in that comorbidity clusters for a district. The darker the colour of a cell, the smaller the percentage of deaths in that symptom for the given district while the brighter the colour, the larger is the corresponding percentage.
From the above heat maps, it can be inferred that in each district, the percentage of deceased patients classifying under the clusters of Diabetes and Obesity and Heart Diseases is much higher than the percentage of deceased patients in other co-morbidities clusters. Thus, the two most important Co-Morbidities Clusters are Diabetes and Obesity and the Heart Disease and Blood Pressure.
