Of coding, COVID & covidiots

Number Of Viewers: PHP Hits Count

Earlier this year, many nations including ours launched a war against COVID 19. We salute all medical and paramedical staff, government servants, transport workers, delivery personnel and all other frontline workers who are bearing the brunt of the crisis.


We are a team of data scientists ( statisticians, computer scientists and economists) who, though lacking such frontline skills, want to use what skills we do possess to contribute towards understanding and monitoring the crisis. Inspired by Napoleon's quote that war is 90% information, we have used information technology to analyse available COVID related data with a focus on India and present our collective results on this page.


Our analysis includes the following:


Many of our model choices have had to be pragmatic in the light of limited information related to the outbreak. We plan to keep refining and updating our analysis . We are open to the idea of establishing research collaborations and hope that this can eventually be turned into a monitoring tool which will aid our preparedness for the future.


We look forward to your reactions and comments.


Table 1a: Top 10 state/UT by Confirmed cases

Table 1b: Top 10 state/UT by Death

Table 1c: Top 10 state/UT by Recovered Cases

The above three tables give top 10 States and Union Territory in terms of Infected as well as Deceased cases. The numbers are reported in per million citizen in the geographical unit. Deeper color indicate higher numbers both in absolute values as well as values per million.

  • In terms of Confirmed cases. Delhi and Ladakh top in terms of numbers per million population ( Ladakh and Andaman Nikobar is not so densely populated)
  • In terms of deaths, Maharashtra is worst affected followed by Delhi
  • In terms of Confirmed cases. Ladakh and Chandigarh top in terms of numbers per million population ( Ladakh, Chandigarh & Jammu Kashmir is not so densely populated)

Figure 1: Comparison Of Different Countries

The above figure shows the daily new cases confirmed in first 32 days.

  • India's growth in number is less as compared to the European Countries
  • Italy and Spain have the steepest of Growth

View more on https://www.kaggle.com

Figure 2: Spread Of Infection across different states

The present analysis tries to point out the states that attributes mostly to the fatal spread of the pandemic in India. Here, the states are being evaluated based on the following 3 rates (Calculated per 100000 exposed):

    • Infection Rate
      Recovery Rate
      Case Fatality Rate
  • View More

    Consequent to the lockdown the migrant labour in different states, mainly in Western and Southern India, had to go through a very difficult time – unemployment, hunger, fear and actual cases of increasing infection. In the recent times probably there is no parallel except in times of war in any country that forced a large section of its own citizens to the level of distress that Indian migrant workers had faced. The distress forced them to leave the place of their livelihood until very recently, often thousands of kilometers on foot. This has been discussed and debated in media and other mediums. The purpose of this write up, however, is not to add another piece on the miseries of the migrant labour during the lockdown period, rather what will happen when lockdown is lifted and economic activities start taking place. The objective of this short write up aims to highlight the important role that the migrant labour plays in the industrial sector of the states with high concentration of industry. View More

    Figure 3 a: Interest Over Time

    Figure 3 b: Top 5 cities in all cases(Zoom, Google Classroom, Skype)

    View more on https://www.kaggle.com

    Figure 4: Mobility Trend vs Time

    Apple has made its mobility data available and Google trends data for search on Coronavirus is available. here is a short analysis on these two for India.

    Figure 5 a: Graph Of Confirmed Cases in India

    The data for this study was downloaded from www.covid19india.org . The following plots show the current COVID-19 incidence in the different Indian states. It also depicts the timeline of incidence of COVID-19 cases in India.View More

    Figure 5 b: Daily Confirmed Cases in India

    The raw data available on the website www.covid19india.org is mostly qualitative in nature. A brief exploration of the same produced the above barplots. Here we see the number of confirmed cases announced and the number of cases that changed in status (either from confirmed to recovered, confirmed to deceased).View More

    Figure 6: Prevalence of COVID-19 in Delhi Population

    As COVID-19 pandemic wreaks havoc across continents, and vigorous public health responses are now being put in place in all the countries hit by the virus, articles are appearing in the the mainstream media explaining the importance of such public health interventions in flatteningthecurve.View More

    Figure 7: Map of Nizamuddin Basti

    In India one such case has been identified at the Nizamuddin basti in Delhi. Nizamuddin is a crowded, busy neighborhood of narrow lanes lined with market stalls and tiny shops, known for two important historic sites. The map shows just how densely populated the locality is.View More

    Figure 8: Comparison of New cases To Existing cases in Different states

    The rate, defined as the ratio of the number of new cases to the number of existing cases, is plotted. The black line is for the states of Maharashtra, Gujarat, Rajasthan and UP. Together, these states make up for 49% of the total cases. The red line is for the rest of the states.View More

    Figure 9: Type Of Sources

    In the wake of Covid-19, all nations have taken certain interventions or measures to safeguard their people and keep the effects of the pandemic at bay. A rich database of Government Measures of 196 distinct countries is being created and maintained by an organization called ACAPS. View More

    Figure 10: Volatility of Stock Price and Structural change in volatility for different airline companies

    On the first of 2020, COVID pandemic hit the world and different countries declare a complete lockdown. As a result of this, the transport system totally collapsed. With no difference, a similar case has been noticed in the aviation industry as different countries banned international travelling from the end of January 2020. Here, we have analysed the stock price of a number of international airline companies and tried to find the impact of the pandemic on this particular industry. We have also estimated the dates when a significant change have been noticed in the stock market and explained the economic significance of those dates. View More

    Figure 11a: Agreement of the observed and model based daily case numbers

    In this analysis we shall fit a compartment model to COVID 19 outbreak data for India in the pre lockdown period.We shall use an extension of the SEIR model which includes the following compartments View More

    Figure 11b: India Corona Virus Affected

    Barring a few exceptions, there is almost a linear relationship between the number of confirmed cases and the number of deaths. Since the virus has affected all the continents across the globe, this would suggest that climate, demography, ethnicity or economy has yet to play any role in the death rate due to the disease. View More

    Figure 12a: Semiparametric regression model to the log case rate per 100000 on time since identification of the first case

    In this analysis, we will identify interventions which have had an impact on the trajectory of the epidemic curve. In particular, we will focus on the various interventions which have been adopted in India.View More

    Figure 12b: Daily infection count and rate in India from 30th Jan

    The spread of Covid-19 in India: infection rate, death rate and recovery rate. What is the impact of lockdown and its relaxation on these? Where is India compared to countries worldwide in terms of curbing the virus trajectory? View More

    Figure 13: (above)Estimated Effective reproduction number of Bihar(Very Low Testing Rate)
    (below)Estimated Effective reproduction number of Delhi(Very High Testing Rate)

    In this study we have estimated the Reproduction Numbers (R0), for the different states of India and observed their trends. We have classified the states into four ordinal clusters : “States with very low testing rates (Bihar, Jharkhand, Uttar Pradesh, West Bengal) ”, “States with moderately low testing rates (Assam, Chandigarh, Haryana, Madhya Pradesh, Odisha, Punjab) ”, “States with moderately high testing rates ( Himachal Pradesh, Karnataka, Tamil Nadu, Uttarakhand )” and “States with very high testing rates ( Delhi, Goa, Gujarat, Jammu and Kashmir, Jammu and Kashmir, Kearala, Maharashtra )”. View More

    Figure 13b: Estimated Reproduction number of westbengal

    In this study we are estimating the Reproduction Numbers (R0), for the different states of India and observed their trends. We have classified the states based on the proportion of Covid19 positive cases among the total samples tested. Usually, for a particular state, more the proportion of positives among the total tests performed, more grave the situation is and more testing is required to be performed.View More

    Figure 13c: Active cases for districts of India over lockdown periods

    The above maps are showing the number of active cases for districts of India over lockdown periods. From lockdown 3.0, the active values are measured on the day on which that phrase of lockdown starts. However, for lockdown 2.0, we have measured the number of active cases as on 26.04.2020 (based on the available data). View More

    Figure 14 a: Daily Infection rate in 1 million people

    The novel coronavirus, COVID-19, originated in China and has spread across the globe very rapidly. As of 6 April 2020, there have been almost ~70000 death and ~1.3 million infection across the world due to the coronavirus pneumonia pandemic( COVID 19).View More

    Figure 14 b: Active cases since first case reported

    The following analysis explored possible downturn of COVID 19 infection rate in India. Daily active cases are taken for the analysis. As daily number of active cases are count data, we converted that into log scale to make it continuous before apply semiparametric regression method to estimate the curve over time.View More

    Figure 15: Lockdown Count Predictions vs Actual cases

    On March 24, 2020, PM of India, Sri Narendra Modi announced an unprecedented nationwide lockdown. Immediately, the economy has taken a huge hit. Daily wage laborers are immediate front line soldiers who took the hit. Also, the travel and hospitality industry is another industry and its people which are in big trouble. Overall, the cycle of the economy is being stopped.View More

    Figure 16: WordCloud Of Tweets

    This is prepared by crawling tweets from Twitter containing keywords like - "#Covid_19", "#coronavirusindia", "#CoronaChainScare", "#StayHomeStaySafe", "#CoronavirusPandemic", "#CoronaVirusUpdate". The tweets corpus consisted of 41,402 tweets which were crawled from 27th February, 2020 to 30th March, 2020. The field consists by following fields of information- tweet content, creation date, screen name and retweet count.View More

    Figure 17: Effect Of the pandemic on Nifty Stock Closing Value

    The first case of the 2019–20 coronavirus pandemic in India was reported on 30th January 2020, originating from China. The outbreak has been declared as pandemic by world health organization (WHO). Here we are finding the effect of the pandemic on various economic indicators such as View More

    Figure 18: Stock Log_Return Of Nifty 50

    The first case of coronavirus pandemic in India was reported on 30th January 2020, originating from China. The outbreak has been declared as pandemic by World Health Organization (WHO). Here we have collected stock data from different sectors and split data into different phases. Nifty-50 is used as stock index. View More

    Figure 19: India Local Level BSTS, Actual vs Predicted

    To understand the behavior of Covid-19 cases’ time series data for different countries, Bayesian Structural Time Series models were applied to get some insight on forecasting behavior for different variants of the models.View More

    Figure 19b: Actual vs Predicted Deaths per population(100000) in India

    To understand the behavior of Covid-19 cases’ time series data for different countries, Bayesian Structural Time Series models were applied to get some insight on forecasting behavior for different variants of the models.View More