Thread: After more than a year of work, we have now launched , a public platform that seeks to expand access to Indian data for everyone, and deepen the understanding of the country that can come from this datadataforindia.com We take large public datasets, make them usable, derive descriptions and insights about the country from them, visualise these insights through beautiful charts, and make all of it freely available - the insights, the charts and the underlying data.

Thread: For an article I wrote for @Article14live, @P39A_nlud gave me access to a trove of data on 88 people on death row, one-fifth of India's death row population at the time. This included demographic data & transcripts of long interviews with the prisoners and their families. I also had access to diagnostic data on their cognitive abilities, intellectual disabilities and mental disorders - data that has never before been collected on this population, after tests that the people on death row consented to participate in.

It's hard to overstate how significant this is - covid19india.org will cease operations at the end of October. They've explained their reasons here blog.covid19india.org/2021/08/07/end/ I'm not sure how many people realise that India - one of the countries the worst affected by covid and a tech powerhouse - has no official website for detailed covid data. A volunteer-driven crowd-sourced website has been the source for everyone (including govt at times).

A story of Indian data in four acts:
Act I: hearing of the trouble people were having in accessing routine health services during the first lockdown, I looked up National Health Mission (NHM) Health Management Information System (HMIS) data and wrote this livemint.com/news/india/how… Act II: when the next round of data came out, I found that the disruption had been even worse than we had known and wrote this in August 2020. The HMIS website then stopped publishing monthly data indiaspend.com/covid-19-disru…

Thread: last week, I reported on data from the National Health Mission's administrative data portal that showed a sharp spike in mortality in April-May 2021. This is official data from over 2 lakh health facilities, mainly rural and govt-run indiaspend.com/covid-19/death… The report noted the increase in deaths from "fever" and "unknown causes" in particular during the second wave's surge

My heart sank when I saw these numbers: new official data from the National Health Mission - compiled mainly from ANMs and PHCs ie rural areas - show a big spike in deaths recorded all-India in May 2021 indiaspend.com/covid-19/death… The NHM captures less than a third of all deaths in India, since its focus in rural areas and govt facilities. But within this limited universe, there was a surge in deaths from "fever" and "unknown causes" in May 2021

For @the_hindu I have this primer on using the Civil Registration System to estimate excess and pandemic mortality in India. A few thoughts: thehindu.com/opinion/op-ed/… What we're trying to do with the CRS (and what the WHO, people in several other countries are trying to do) is produce estimates of excess mortality from which experts can estimate the extent of missed covid deaths. This comes from accepting that many covid deaths will be missed.

For @IndiaSpend I looked at excess mortality data in five states to think about how this alters India's official death toll from covid. The first thing is that the ratio of the official covid toll to excess deaths varies from 2% in MP to 60% in Kerala for 2021 The second thing to think about then is how many of these "remaining" excess deaths are also covid deaths. Without good cause of death data, it's hard to know. But even with cause of death data, as in TN, there's going to be misclassification.

Here we go again: today, I am reporting all-cause mortality for Andhra Pradesh (population: 53 million). In May 2021, AP saw nearly five times as many deaths as the state usually sees in May scroll.in/article/997427… In all, between January 1 and May 31, 2021, AP saw 130,000 excess deaths (compared to the 2018-19 average). Not all of these will have been deaths from covid, but for perspective, the excess mortality was 34 times the official toll from covid.

Some updates to my thread from yesterday on elevated all-cause mortality in Madhya Pradesh

https://twitter.com/Rukmini/status/1403561579577151498

As promised, here's the link to my GitHub including monthly totals from Jan 2018-May 2021, and districts data for the months of April and May (2018-2021) for MP. I'll update it with more data in the coming days/ weeks github.com/Rukmini-S/MP-d…

Thread: Today I am reporting official all-cause mortality from the state of Madhya Pradesh - with 84.5 million people, it is India’s 5th most populous state. It was also India’s fifth poorest state as of 2020. It is predominantly rural, with just 29% of its population in cities. I wrote this for @DainikBhaskar and here's the headline: MP saw twice its usual (2018-19 average) deaths in April 2021, and *five times* the usual in May 2021

Thread: I am supposed to be taking a break from regular reporting to finish up my book. But I cannot let it go when the government puts out false or misleading data. When I saw this in the April 21 press conference, I knew there was a problem

https://twitter.com/COVIDNewsByMIB/status/1384829869255655428

It seemed highly unlikely to me that the government was either able to track every person who had received even one dose of a vaccine, or that every person being tested was being asked their vaccination status. So I asked people, labs, and officials. And that data's faulty.

There should never have been a contempt of court case against Prashant Bhushan, a great defender of our democracy, and his conviction is a travesty. But this Re 1 fine and the joking around it rankles. In 2011 I stood in a Saket court to watch a 19-year old boy's trial. (thread) He had been picked on charges of stealing a wallet containing Rs 200. For one entire year he remained in Tihar after being granted bail because he did not have the money for a bail bond. His final escape? Conviction. Because the jail term was 3 months.

For @livemint, I looked at recent population projections for Indian states for the next 15 years, and I was struck by the generational divide that is coming up between India's north and south. Short thread follows: Fifteen years from now, the median Tamil person will be over 40 while the median Bihari will be not yet 30. That's a considerable age gap, and it has widened further.

Bhutan went into national lockdown yesterday following one young woman testing positive. Short thread on how a woman who followed all the rules still triggered a national crisis, and how the country is containing it. The 27-year-old woman returned from Kuwait on June 26 and was placed in mandatory institutional quarantine. She was in quarantine for one month (!) over which time she repeatedly tested negative on RT-PCR tests but positive on antibody tests. On July 26 she was released.

Here's a new problem for India's testing data: states like Delhi are conducting both RT-PCR tests and antigen tests, but reporting positive test numbers combined. Antigen tests have much more false negatives, so start taking falling test positivity rates with a big lump of salt. I wrote about it here for @IndiaToday: indiatoday.in/diu/story/why-…

Little thread: For work, I've been looking up other countries' coronavirus dashboards, and one thing struck me about Brazil's (aside from the fact rhta it's very good). Look at that second tab. SRAG is "síndrome respiratória aguda grave" - or as we know it in India, SARI or Severe Acute Respiratory Illness

(Thread) I looked at data from 627 districts and over 1.5 lakh health facilities for the last few months and years, and what I found was very worrying. Medical services declined across the board in March 2020. This includes immunisations - 2 lakh children fewer got the pentavalent vaccine, over 1 lakh fewer kids got their BCG vaccine. It also includes a host of maternal services.

For @livemint I looked at the data on the healthcare-seeking behaviour of Indians to understand what we need to overcome in the fight against the spread of the novel coronavirus. Thread: The poorest and most marginalised are most likely to report ill-health especially - and this is worrying in the current context - coughs and fevers

I'm trying something new. Every night I'm doing a five minute mini podcast on one question around the novel coronavirus in India. Yesterday's was on how long lockdowns might need to be. You can listen to The Moving Curve - Episode 1 here:
medium.com/@rukminiwrites…
Sound: @mdeii Episode 2 of The Moving Curve (my experiment with a 5-minute daily podcast on the novel coronavirus in India) considers how strict a lockdown needs to be to be effective
medium.com/@rukminiwrites…
Sound: @mdeii

Thread: For @ThePrintIndia I looked at the current rate of growth of confirmed COVID-19 cases in India and estimates by others. In most scenarios it is looking like India will hit 1 mn *confirmed* cases in May - and the fatality rate is broadly over 3% theprint.in/opinion/curren… Here's a simple extension of the 5-6 day doubling rate, and here is a detailed projection by a team of bio-statisticians.