Garg was a addressing a session at AIMA’s 53rd National Management Convention here.
A harmonised data infrastructure is extremely necessary to ensure that data can be systematically discovered, understood, connected and utilised, said Secretary, Ministry of Statistics and Programme Implementation, noting that at times two words have different definitions.
He said the central as well as state governments are involved in a five-step process to build such infrastructure, especially given the fact that data is the raw material for AI.
“Five-step process that we have started both across ministries and departments and states…the first step is to have government data which is harmonised and then we can look at other data in the private sector and others can join in,” Garg said.
The second step is maintaining metadata, he added.
The next will me bringing uniformity in definitions across ministries, departments.The fourth is to have identifiers so that different data sets can be compared. And the fifth step towards a harmonised data infrastructure is having a quality assessment while using data.
Garg said the availability of updated data is necessary for the goal of Viksit Bharat by 2047. To ensure data harmonisation, it is important that the quality of data is universally recognised.