2022
The Hindu’s coverage of excess deaths during the COVID19 pandemic in India
Country/area: India
Organisation: The Hindu
Organisation size: Big
Publication date: 28/05/2021
Credit: Srinivasan Ramani, Vignesh Radhakrishnan, Pratap Vardhan
Biography:
Srinivasan Ramani is the deputy national editor with The Hindu newspaper. He anchors the opinion pages of the newspaper and is also in charge of the explainer journalism (“Text & Context” pages in the epaper) and data journalism sections. A journalist for a decade and a half, he has a PhD in Comparative politics from the Jawaharlal Nehru University besides an engineering degree and experience as a programmer analyst with Infosys Technologies Limited early in his career.
He leads three exclusive (and small) teams in The Hindu for the roles mentioned but they also work in some coordination. The oped page for e.g. publishes a data-driven story in the form of infographics and text every week-day and the Text & Context epaper publishes a full page data-driven graphical explainer every fortnight.
The Hindu Data Team (Srinivasan Ramani, Vignesh Radhakrishnan, Sumant Sen & Jasmin Nihalani) also helps with graphic value-adds and investigative stories based on data for the print and online editions. It has produced hundreds of “Data point” stories in the oped section, several full page explanatory stories using graphics and data visualisation on various matters of public import (coverage of elections, yearly national budget, disasters, stories on political economy, releases by surveys etc).
The excess deaths project was anchored by Srinivasan Ramani with contributions by Vignesh Radhakrishnan and an external hand, Pratap Vardhan for coding support. The team also collaborates extensively with The Hindu’s Design team and these collaborations have received international recognition (Society of Newspaper Design Awards in 2017 for e.g.).
The Hindu Data Team also dedicated itself to tracking the spread of the novel coronavirus in India by maintaining a few trackers (cases & deaths, vaccination progress at thehindu.com/coronavirus/ and regular graphics on print), comparative exercises with other countries, impact of the pandemic on the economy etc in the form of several stories during the period as well.
Project description:
I anchored a series of data-driven stories on assessing the “excess deaths” during the COVID19 pandemic in India. Using the Right to Information Act, web scraping and reporting tools, I along with other colleagues across several bureaus in the newspaper I work in, The Hindu, managed to extract data on d from the Civil Registration System (CRS) for nine States (provinces) and five cities. We used three models to estimate the “excess deaths” during the pandemic and went on to assess that excess deaths were close to six times the official count of COVID19 deaths for those States in the
Impact reached:
There was widespread anecdotal reporting on a sudden increase in deaths across States in India during the second COVID19 beginning March 2021 in particular. Our reporting on excess deaths using the CRS data was valuable in placing the true picture of the devastation wreaked by the pandemic. The figures released by The Hindu in its “excess deaths” coverage were used for peer reviewed research by demographers, epidemiologists and public health experts. This also resulted in a public debate over the actual COVID19 death toll, that culminated in greater allocation for compensation for deaths by some States.
Our work was also useful in bringing together a more thorough understanding of the impact of the COVID19 pandemic as several publications used our base data to assess excess deaths in India. See for e.g. – https://www.economist.com/graphic-detail/coronavirus-excess-deaths-estimates. Or this – https://www.science.org/doi/10.1126/science.abm5154. This also clearly established that India was no exception (as had been argued by some during the early impact of the pandemic) in terms of how the pandemic affected lives and the high multiple of excess deaths in a country with some States having poor primary health care and infrastructure revealed that the impact was not different from other countries with similar socio-economic profiles (such as Latin America).
Our work was also vital as India’s civil registration system on deaths is not uniformly maintained or publishes data in real time unlike countries in the developed world. By gleaning information State-by-State and city-by-city, we managed to pierce together a truer extent of the mortality during the pandemic that brought the mortality data in line with experiences elsewhere. See this summed up story for e.g. – https://www.thehindu.com/data/excess-deaths-during-the-pandemic-in-india-was-58-times-the-official-covid-19-death-toll/article36405310.ece
Techniques/technologies used:
Gleaning data from the Civil Registration System required us to utilise several tools – 1) The use of the Right to Information Act to procure data from civic authorities and public health officials in governments of several States and cities (Karnataka, Kerala, Himachal Pradesh, Punjab, Delhi among States/ Union Territories and Thiruvananthapuram and Chandigarh among cities). Some State officials were prompt in responding to these queries, some took their time. We also pierced together data for the capital region (Delhi) by filing separate RTI requests for each of the five municipal corporations within the region. 2) For some States such as Rajasthan, Haryana, West Bengal, Maharashtra and cities such as Hyderabad, Kolkata and Chennai, data on deaths and certificates were available on the web. For data that was available on State websites, we used web scraping techniques to scour data from the app by aggregating them through requests using python scripts. Some data such as those for Hyderabad and Rajasthan were available in Android apps. We used screenshot text grabbing software to download data and manually aggregated and fed them into spreadsheets. For Rajasthan’s data, we used Android app simulations on the computer and used them to scrape data and aggregate them. 3) Data aggregated into spreadsheets were then parsed month-wise for the period April 2020 to May 2021 (the beginning of deaths due to COVID19 till the peak of the second COVID19 wave driven largely by the Delta variant in India). Using three models to estimate base mortality (average of previous five years, for the same months, for States that had a steady number of deaths, average of previous two years, for the same months and extrapolating death estimates using a linear increase), we calculated the excess deaths during the pandemic period by comparing the estimates with actual death
What was the hardest part of this project?
The hardest part of the project was sourcing the data. Civil Registration Systems in India vary in implementation across States – data in some States are fully digitised, in some, they are partially digitised and in others, not at all. Then there are variances in registering/reporting deaths to civic authorities with some States such as Bihar and Uttar Pradesh having low registration numbers. The use of the Right to Information Act was useful in States which were pro-active in sharing information and where the data was maintained well in digital formats.
Where this was not so, we had to rely on web scraping techniques to parse available digital data and come with estimations for the overall data by looking at prior Civil Registration System reports released by the Registrar General of India which analysed the registration levels, delays in registration etc in each State.
Web scraping was relatively easy to utilise when structured data in queryable form was available in States/ municipal corporation websites. In cases where this was available in apps, it required a more concerted effort and involved coding to write scrapers and to parse queries to download structured data.
We seek a recognition of this effort because of the exclusivity of the work (while other journalists/outlets also worked on excess deaths projects in two or three States each, The Hindu managed to produce excess deaths information from the highest number of States and cities), and the utilisation of various tools at the disposal for journalists – the Right to Information Act, web scraping and sources. The jury could also look at several published and pre-print papers on excess deaths estimations across several countries that have further utilised our work – https://www.science.org/doi/10.1126/science.abm5154 , https://www.medrxiv.org/content/10.1101/2021.09.30.21264376v1 , https://www.medrxiv.org/content/10.1101/2021.08.04.21261604v2 for e.g.
What can others learn from this project?
Good journalism is always an outcome of collaborative efforts. To produce the stories on excess deaths – most of which are archived here – https://www.thehindu.com/specials/coronavirus-spike-in-excess-deaths-during-pandemic/article34876318.ece, we had to consult demographers and epidemiologists to understand the methodology of utilising the Civil Registration System and to calculate excess deaths. To source data, I had to work with reporters from several bureaus who helped in filing Right to Information Act (RTI) requests at respective municipal corporations and State health departments. Wherever these RTI requests were sourced by help from reporting colleagues, they were given joint bylines in the analyses on excess deaths for those States/ cities in the published version of the article. Data journalism is also a subset of investigative and analytical journalism. The knowledge on how to go about using the “excess deaths” methodology was provided by oped writers (demographers and public health specialists) for The Hindu (I happen to manage the opinion section in the paper). Without their guidance and nudging, the project woud not have been possible. Once we understood what we required to do for each State/ city and what we needed to process as data, we went about utilising the tools of data journalism to proceed with the project. In sum, it takes a village to come up with effective investigative stories in a newspaper and data journalism, sitting on the shoulders of expertise from publicly minded intellectuals and citizens in civil society, is most effective done that way. This is the take-away of our work, in my opinion. Data provides us a vital instrument in documenting the effects of a pandemic and to construct public health responses to it. The excess deaths project was part of a series of data driven stories on the spread of the coronavirus in the country (The Hindu’s data team also
