BS Number Wise: Govt's data website is a hard ask for information

Portal data.gov.in is difficult to negotiate and it is outdated, defeating the purpose of understanding economy and policy

The Union Budget this week projected the government’s balance sheet for 2022-23, besides giving revised estimates for 2021-22 and actuals for 2020-21. It listed the government’s plans for the financial year starting April; accompanying documents gave receipts and expenditure along with the outcome of schemes. What the government did not give is a time series.

For instance, the Budget said the government plans to spend Rs 1 trillion in 2022-23 on education. It did not tell what the government spent, say five years ago or two decades ago. It also did not tell how spending on education has been in a decade as a proportion of the GDP.

One has to collect data from Budgets over the years to determine such spending—a tedious task.

Not just the Budget, most annual reports by ministries do not present time-series data. To fix this and encourage research, the government announced an open data platform in 2010. The portal, data.gov.in, was to mirror other such initiatives across the world. The US parallel, data.gov, presents historical data on department-wise expenditures.

A decade after its launch, data.gov.in still does not meet demands for data. The website is outdated and its Application Programming Interface (APIs)—meant for easy access and automatic updating of data—and datasets are rarely updated.

The website segments data according to sectors and catalogues, but it is impossible to navigate it and extract information.

Except for Tamil Nadu, no other state had put up its Covid-19 medical bulletin in seven days. Of the 72 most recent entries on different data sources, a majority belonged to the southern states. Accessing these 72 entries was not easy in most cases, as the website would redirect to state websites. Not all states have linked their websites and neither have all government departments. Tamil Nadu led states in terms of data initiatives. On the other hand, the Smart Cities mission formed a sizeable chunk of central government’s data on data.gov.in.

The problem in accessing data does not end there. Analysis shows that 85 per cent of the APIs were for just three of 35 sectors defined on the website. Health accounted for 52.5 per cent, followed by the national census (21.4 per cent) and agriculture (11.4 per cent).

The three sectors dominated in the availability of datasets as well. Of the over half a million datasets available on data.gov.in, 60 per cent were for the health sector; agriculture accounted for 20.5 per cent share and census 7.1 per cent.

Some states uploaded their Census 2011 data in 2021. In the Parliament category, the last update was in 2018 and data was only available till 2014.

In terms of data visualisations, only 229 were available for eight of 35 sectors. Despite being the biggest contributor to the website, Health did not have a single data visualisation available. A third or 81 visualisations were for the industries sector.

Ministries' annual reports, department websites and the Budget contain troves of data that can help in understanding the economy better. The government needs to bring all this information to one accessible website.