About the webpage
About the dashboard
This page of the site is developed to share a presentation of various data sources covering COVID-19 issues. This is a static form of dashboard, prepared in form of web page. The dashboard is designed to give a sense of how fast the COVID-19 pendemic is progressing. The dashbord is not a complete presentation on the issue of COVID-19 and is actually inspired various many activities presented by individuals and institutions, who gave an excellent understanding and intrepretation related to current state of observations on COVID-19. Many of these sources are acnkowledged in their respectives places in the dashboard. I greatly admire these efforts, without which this work could be possible.
The main objective of the dashboard is to present the COVID-19 cases and the pregression of the Pendamic in India. Since the pendemic is in progression, any interpretation of the data is not final because many new understaning is unfolding due to the variablity of data on a day to day basis. I intend to update the page as per the changes in the data content, however, this is subjected to quality and accessibility of the data available in the public domain. In presening the case of COVID-19 progression in India, this dashboard aims to describe the district level COVID-19 situation in India.
Content of the dashboard
The dashboard is created in form of web-pages. These web-pages can be explored by browsing the corresponding tabs on the top. The content of the tabs can be explored using the pop-up menu given in the corresponding places.The descripiton of tabs are:
1.About- this is the current tab, it has two pages About the web-page (the current page) and acknowledgements.
2.Global Scenario- this page shows current status of COVI-19 data. It presents the visualization to enhance our understanding and interpretation of the data. This has two pages (i) a visual of represntation of key observation of current COVID-19 data in form of treemap and bar chart, (ii) a graph showing the relationship between testing and COVID-19 cases.
3.Indian Context- this page shows current status of COVI-19 data in the Indian context. The first page of the tab presents a description of COVID-19 cases in India, the second page attems to draw some inferences from the status of COVID-19 cases and the role of testing in differnt countries. The page attmpets to extrapolate these inferences in the Indian context to explore the expected COVID-19 positive cases, if corrected with the testing background inferred from the relatioship between COVID cases and testings. The third page shows the mathematical approach for predicting COVID cases in India. The page explores the possible dates for peaking and suppression of COVID cases in India.
Information about the author
The content of the dashboard-cum-wepage is prepared by Kamal Kumar Murari.
Primary Affiliation: Centre for Climate Change and Sustainability Studies, School of Habitat Studies, Tata Institure of Social Sciences, Mumbai.
Secondary Affiliation: Labour Market Research Facility(LMRF), School of Management and Labour Studies, Tata Institure of Social Sciences, Mumbai.
In LMRF, I am the core faculty of Executive Post Graduate Diploma in Analystics (EPGDA). This work is an outcome of teaching and research activities in EPGDA.
Disclaimers: This page of the blog presents an independent work, not funded by any individual of agency, and not associated to any institution (public or private). The work presented in the web-page are personal and are not associated (either in form of observatoion or interpretation) to any governmental, non-governmental and polictical party. At the moment, when COVID-19 crisis is posing a great danger to humanity, the work in the blog-post attemps to provide a lead in terms of our understanding of how the crisis will unfold in the future. The content presented in the web-page is strickly for the education and research purpose and do not claim for the accuracy of the predictions. I use the predictive modeling approach to provide a background on the pattern of COVID-cases, peak cases, peak date, and end dates of he crisis based on certain assumptions. At the moment, there is no clear background information and evidence to validify these assumptions. Readers must take the observation and interpretation with caution.
World Corona Cases
COVID-19 cases and deaths
Treemap of World Corona Cases
This graph shows the treemap of current state of Corona cases in the world
Notes:
1.The size of the blocks in chart A refers to the relative share of total number of positive Covid-19 cases in the world
2.The color represents the number of deaths for individual countires
3.Twenty five countries cover more than 85% of the total positive Covid-19 cases in the world
Death and Recovery rates of selected countries
The Figure below shows the status of total death, recovery and active cases for selected countries of the world
Notes:
1.Death and Recovery rates are with respect to total Covid-19 cases
2.The numbers of persons recovered are missing for Netherland and UK
2.Active cases referes to cases currently active as percentage of positive Covid-19 cases
Testing and Corona Cases
Testing and COVID-19 cases
Corona Testing and Cases (for top 25 countries)
Notes:
1.The figure is for selected countries that has more than 85% of the global Covid-19 cases
2.Dashed red line indicate the trend line between testing per million population and Covid-19 cases per million population of these selected countries
Corona Testing and Cases (for Asian Countries)
Notes:
1.The figure is for Asian countries
2.Dashed red line indicate the trend line between testing per million population and Covid-19 cases per million population of these selected countries
Covid-19 Cases in India
Testing for COVID-19: Indian context
Policy Response to Corona Crisis
India reports less number of positive Covid-19 cases compared to many countries in the world. Interestingly, time series data indicate a slower rate of rising of Corona cases in India. Much of the credit for the slower rise in Corona cases goes to the timely decision and strict implementation of nationwide lockdown. However, the implementation of lockdown has serios implications, particulrly on poor and vulnerable section of the population. Most effected are the poor people living in the bigger cities, who got no ways to hadle the adversities brought out by the lockdown period. Till May 3, the country was under strict lockdown for about 40 days. This has strong negative imlications on economy, employement, and welfare of India’s population.
Indian government announced the third stage of countrywide lockdown from May 4 to May 18. The third stage of lockdown will have some conditional relaxation depending upon the location and history of Corona cases. Is lockdown a right policy measure to address the COVID-19 crisis? This is an importnat question which policy makers and acadmecians are focussing on. Ferguson et al. (2020) suggested that there are two ways of dealing the COVID-19 crisis: (i) mitigation, that focuses on policy strategies towards slowing of the pendemic but not necessirily stop its spread. This aims to reduce the healthcare demand, and protact the vunerables to the disease infection. Lockdown, social distencing and quarentine are the kind of policy responses that focus on mitigation, and (ii) suppression- which aims to reverse the pendemic growth, reduce the case growth to the lower level and maintain the situation for an indefinite time. This can be achived by contact tracing.
The World Heath Organization also pointed out that evidence from the experience of Wuhan indicated there are three key types of Corona carriers- (i) the symptomatic carrier- individual who show definite symptoms related to COVID-19 infection, (ii) the pre-symptomatic carriers- who do not show any symptom, at the moment, related to the COVID-19 but has high probability in coming to become symptomatic, and (iii) asymptomatic carriers- who neighter show any symptom nor indicate any possibility of turning to be a symptmatic. identification symptomatic carriers is more easier than the other two that does not require mass scale contact tracing. Whereas, identification of pre-synpromatic and asymptomatic is only be done using testing. At the moment, there are a number of approaches of testing. Testing of popolation is one of the important policy response to address the COVID-19 crisis. In absence of mass scale testing, any numbers related to positive exposure of COVID-19 is a gross underestimation, without testing it is hard to reach out at the state of supression of pendemic as pointed out by Ferguson et al. (2020). This is a serious concern for all countries, particularly where the spread of the infection has reached at local and community level. For countries, where are the heath infrastructure is poor- such as India- this is even more serious because without mass scale testing the lockdown will only delay the danger. Lockdown will only delay the possibile burden on heatcare infrastructure, but will have a serious repercussions on the economic progress. Longer implementation of lockdown will impact all section of population but seriously impact the poor and vulnerables. Therefore, as recommended by Ferguson et al. (2020), mitigation and supression both shold go side by side for the proper addressal of the crisis. Testing is one of the important policy response in this regard.
Tracing Corona Cases in India
The Coronavirus crisis is unprecedented challnge for the world. The policy response, decision making and implementation to address the crisis is highly debated debated nationally as well as globaly. Testing is recommended to be the most effective response and widely accepted by the policy makers globaly. However, many countries, particularly developing and least developing nations, are not equipped to conduct mass scale testing. Figure 1, shows the relationship of testing and the reporting of positive COVID-19 cases. I took both testing and COVID-cases per million of population for standerdization so that I can compare the countries. In the Figure 1, the horizontal and the vertical scale is represented in a logrithmic scale to accomodate all countries because of very high dispersion in tests and COVID-19 cases among countries.
In the previous tabs, I have plotted two figures to show the relationship of testing and positive COVID-19 cases for selected countries. The Figure 1 is a combined form of the two figures of the previous tab. Clearly on log-scale, there is a strong linear relationship between testing and positive COVID-19 case. This relationship is true, though with different linear parameters, for other data subsets such as for top 25 countries and Asian countries.
Figure 1: The relationship between testing and Covid-19 cases
Notes:
1.The markers for the countries shown in text is highlighted as yellow
2.The trend lines in the graph shows the linear relationship, on log-scale, between Test per million population and Cases per million population for different data sets. The green line shows the trend for G25 (top 25 countries that has positive Covid-19 cases), the red line shows the trend of Asian countries and the black line shows the trend of all countries.
An assessment of COVID-19 cases in India-using linear models
Testing rate in India is very slow and is not comparable to its peer group such as top 25 nations reporting high COVID-19 cases, and Asian countries. At present, India’s testing rate is not comparable to Thailand and Nepal. The plot between testing and cases, for per million population, indicated that there is a strong linear relatinship between the testing and the positive Covid-19 cases.Based on the evidence gathered from Figure 1, I assumer four type of models to describe the relationship between testing and COVID-19 cases. Table 1 shows the description of the model parameters as well as estimation of positive COVID-19 cases in India based on prediction of corresponding models. The estimation of COVID-19 cases in India is based on current testing capacity (which is 984 as on May 9, 2020).
Table 1: Model parameters and COVID-19 projections for India
| Model | Deg.of freedom | Intercept | Slope | R-square | COVID-19 estimation | Remarks |
|---|---|---|---|---|---|---|
| 1 | 180 | -2.49 (0.44) | 0.92 (0.05) | 0.65 | 63684 | Model with all countries data |
| 2 | 22 | -0.65 (0.9) | 0.84 (0.09) | 0.78 | 229779 | Model with countries that have top 25 COVID-19 reporting |
| 3 | 44 | -2.13 (0.71) | 0.88 (0.08) | 0.72 | 68021 | Model with Asian countries |
| 4 | 175 | -0.26 (-0.26) | 0.82 (0.06) | 0.71 | 41852 | Model with all countries but taking fixed effect for Asian countries |
Notes:
1.Parameters for model 1 have strong statsitical significance.
2.The intercept for model 2 has a weak statistical significance.
3.The intercept for model 4 is statistically non-significant for Asian countries.
An estimation of COVID-19 cases in India with higher testing
As discussed in the previous section that India lags with respect Asian and other countires in terms of testing to identify COVID-19 cases. Although, in the recent days we have seen a significant progress in testing facilities in differnt parts, however, still more progress is needed in identification and isolation of potential carriers both symptomatic and asymptomatic. In order to assess the size of COVID-19 cases in India with the increase of testing facilities, I have taken some benchmarks and fitted them into models described above.
I am taking two benchmarks Nepal and Thailand to represent countries in the same region. Iran is taken as representive because the testing numbers in Iran represents the median value of testing of all nations. Two countries USA and Germany are taken to represent an aggrassive strategy of testing. Table 2 shows the a description of these countries
Table 2: A description of selected countries testing and COVID-19 cases
| Country | Testing Nos. | Covid-19 cases | Total Death | Total Recovered |
|---|---|---|---|---|
| India | 984 | 38 | 1787 | 15331 |
| USA | 24186 | 3816 | 74809 | 213109 |
| Germany | 32891 | 2007 | 7275 | 139900 |
| Iran | 6325 | 1210 | 6418 | 81587 |
| Thailand | 3264 | 43 | 55 | 2772 |
| Nepal | 2302 | 3 | NA | 22 |
Note that model 2 and 4 are not statistically strong to take the projection as shown in the Table 1. But for the illustration purpose, I am keeping these models for extrapolating potential COVID-19 cases in India. Table 3 shows the potential COVID-19 cases for the situation of increased testing facilities in India, based on the predictions obtained from the four models. Note in the Table, I am presenting only the maximum likelihood estimation.
Table 3: Estimation of COVID-19 in India with increased testing
| Representative Country | Model 1 | Model 2 | Model 3 | Model 4 |
|---|---|---|---|---|
| USA | 1211750 | 3387175 | 1129850 | 585324 |
| Germany | 1607869 | 4385627 | 1479767 | 754044 |
| Iran | 352764 | 1097354 | 348184 | 193855 |
| Thailand | 191932 | 629381 | 194833 | 112400 |
| Nepal | 139196 | 469335 | 143409 | 84300 |
References
Ferguson et al. (2020)Impact of non-pharmaceutical interventions (NPIs)to reduce COVID-19 mortality and healthcare demand.Imerial College Response Team. DOI: https://doi.org/10.25561/77482.
COVID-19 in India
Progression of COVID-19 cases in India
The Figure below shows a descripiton of progression of the COVID cases in India. The datset for the figure is taken from dashboard page of the Centre for System Science and Engineering at John Hopkins University, tracking daily records of confirmed, deaths and recovered cases due to COVID-19. The data can be accessed from https://github.com/CSSEGISandData/COVID-19
Figure 1: The progression of COVI-19 cases in India
Figure 2 Statewise COVID-19 confirmed cases and growth
Figure 3 Statewise COVID-19 Recovered cases and growth
Figure 4: Statewise COVID-19 Deceased cases and growth
COVID-19 predictions
Prediction of COVID-19 in India
The current COVID-19 crisis is unprecedented in nature that has impacted almost all countries in the world.The crisis is not only a concern for the public health, but has implications in all form of economic activities and therby lives and livelihoods. The manner in which the crisis is unfolding, it is hard get any clue regarding what would be total number infected, when will it end in the future. Serious questions that needs to be be answered are the size if health infrastructure needed to dealth with searious and critical cases, policy respense to deal with mitigation and suppression fo the crisis, identification of strategies that will help in terms of dealing with the negative consequences that emerged due to governmental response to deal with the crisis. Policymakers, acadmecians, researchers are using various modling platform to understand the nature of the crisis and answers the question posed above. The framework of mathematica modeling is widely used to address how the crisis will unfold in the future.
There are a number of published work exploring the COVID-19 data profile. Most of the published work have come from experience of China in terms of dealing with the crisis. As the crisis grappled with the larger geographical area of the world, more data-driven modeling is also coming up both in published and un-published domain. Visit the site (https://didi.edu.sg) for a detailed descripition of various data-driven modeling approaches and their applciability of recent COVID-19 data. In this article, I am using a curve fitting approach to predict the profile of COVID-19 behaviour for the selected countries in the world.
Theoretical background
There are various data repositories proving updated information about the recent COVID-19 numbers. The most popular is the data dashboard and data repository maintined by the Johns Hopkins University Center for Systems Science and Engineering (JHU CSSE). The updated dataset can be accessed from (https://github.com/CSSEGISandData/COVID-19). A quick plot of COVID-19 cases for different countries reveal that the data follows a typical pattern of slow growth in the begening phase, this is following by a very fast increase in number of cases (see country wide plots of confirmed cases in https://www.worldometers.info/coronavirus/#countries) . It is expected that the effort of global and national scale public health responses will result in dampaning effect on the COVID cases and beyong certain time, the number of COVID-19 cases will start decreasing. This behaviour of COVID-19 cases can be well explained by a family of Sigmoid fucntion. The most popular member among the family of Sigmoid function is logistic function. The logistics fucntion has two parts, the growth phase and the revershal phase. For most countries the logistic fucntion well represents the observed pattern during the growth pahse. However, the sigmoid fuction fails to represent the peak time as well as the revershal phase, particularly for countries where the pattern of COVID-19 confirmed cases are still in the growth phase. In such cases, the logistics fuction forces data to consider the current peak as the peak of the cases. Thus underestimates the peak time and the revershal phase of the curve. The other challenge for modeling COVID-19 is unavailbilty of information related to carries, particulalry pre-symptmatic and asymptomatic, which has a potential to change the profile of the COVID-19 confirmed cases in the future. This is a major issue, absence of considering the information regarding potential carrries might grossly understaimate our understanding about the flattening of curve as well as projecting the time when the crisis will end.
In the first part of modeling of COVID-19 progression in India. I am using a simple logistic curve widely used in population ecology to define the growth of a population, but with different connotation fitting it into a description of COVID-19 profile. The framework for the logistics function can be given as:
\[ \displaystyle \frac{dP}{dt}=rP-\frac{rP^2}{K}, \]where \(P\) represents number of COVID-19 cases, \(t\) reoresents days, \(r\) is growth rate on confirmed cases. Ideally, \(r\) should vary with time as observed data indicate, but for simplicty I am using fixed \(r\). The term \(K\) represnts the maximum value of possible COVID-19 cases. The solultion of the above equation is:
\[ P(t)=\frac{K}{1+(\frac{K}{P_0}-1)e^{-rt}}\]This framework for the logistic fuction is taken from population studies where it is assumed the population rate at a time \(t\) is propotional to the growth rate, but there is a limit to the growth beyond which the population growth will be stablized. Therefore the framework has two terms, the first term represents the growth and the second term represents the competetion to growth which will force the growth phase to stablised. See https://www.nature.com/scitable/knowledge/library/how-populations-grow-the-exponential-and-logistic-13240157/ for a complete descrition of the framework. In population studies, \(K\) represents the carrying capacity of a system beyond growth phase will reverse. The time at which \(P\) reaches \(K\) refers as time of peak \(t_p\). Identification of an aprropriate \(K\) value is a challnge, particularly when we have evidence only during the growth phase. Especially, in the context of COVID-19 \(K\) represents maximum population which is susceptble to the infection. This is very hard to determine, particulaly in case when you have high probability of pre-symptomatic and asymptomatic carriers for which there is no information available. Absence of information or evidence related to pre-symptomatic and asymptomatic carriers will make the modeling effort weak. Keeping this limitation in mind, I would like to fit the COVID-19 data to the logistic curve so that I can have some understanding about fitness of COVID-19 data for all India at least in the growth phase and can draw an initial understanding of the time of peak as well as the recession phase. Note that the recession phase may not be correct as there is no information regarding the peak of the people infected due to COVID-19. This is a simple excercise, however, will provide intuitive understading abou the progress of COVID-19. Figure 1 shows the results of fitting of the above framework of modeling to the COVID-19 data in India.
Figure 1: The progression CoVID-19 data in India and its fitting into logistics model
Maharashtra
The scenario of COVID-19 in Maharashtra
This page contains the observed scenario of COVID-19 in Maharastra. The page also presents a modeling to assess key features of future progression of COVID-19 cases in the state.
Progression of COVID-19 cases in Maharashtra
Acknowledgements
Acknowledgements
I acknowledge the inputs, feedback and support of individuals and global community that has made the work possible. The blog uses public domain datasets, which is acnkowledged in thier respective places. I do not provide any gaurantee for the accuracy of these datasets.