The Office of the Registrar General and Census Commissioner of India has the mandate to hold the decennial census that started on February 9. As he leads the country into the 15th census, Census Commissioner C Chandramouli spoke to Aloke Tikku about the exercise to count a billion faces and report on the true state of the nation. Excerpts:
What is the status of the Census operation?
This is the second phase of the census. We have already completed the first phase that included numbering houses all over the country. Using that frame, we have sent enumerators with questionnaires.
What does it take to conduct the headcount?
Almost 2.7 million enumerators have to be used. They had to be trained. The material had to be printed and sent across to about 17,000 locations. That also, we have done with a great deal of difficulty.
The first was the procurement of paper, about 12,000 tonnes. High-end presses had to be selected and they had to be given a digital database — which means a particular tehsil may have several languages being spoken. We do the canvassing in 16 languages.
In a tehsil, there may be ‘x’ number of forms which are to be in English, so many in Hindi, so many in Telugu, Urdu. These numbers have to be assessed, and printed according to the numbers the tehsil has asked for. They have to be packed according to that, labelled and then handed over to the department of posts which then transports it to the location. So you can imagine the logistical nightmare that we have to get the right stuff across to the right people.
Does it always end up at the right destination?
Often, one district’s packet goes to the next district. Then, we have to change it, get it back. It is a massive exercise. There are 17,000 distinct locations where they have to be delivered.
What happens next?
The questionnaires will then come to our 17 scanning centres where each and every record would be scanned and logged into a database of images. These images would be read by sophisticated software and based on intelligent character recognition software and we would be able to process the data, and then, tabulate the various tables —an exercise that goes on immediately after the census and it goes on at a 24x7 basis all over the 17 centres.
Is it a technology-driven exercise?
The processing is totally technology-driven. Many people ask why haven’t you done more automation in the collection phase. The major problem is the volume of the operation. Deploying about 27 lakh machines, training operators to use these machines in 16 languages is a hurdle in using the machines in the field.
Other countries that have deployed machines , Brazil or Oman, for example, have less number of languages to contend with. So they are able to use PDAs to collect the data.
But in data processing, we are probably the most sophisticated across the world. We are the first to use ICR intelligent character recognition in last census. This has been further improved this time.
You had used ICR in the 2001 census too. How is it different?
Last time, since it was learning experience for software developers as well as us. We learnt a lot of lessons. The speed of recognition engines has been enhanced, the processing has been enhanced, intelligence has been built into character recognition engines for a better rate. So what used to take eight to nine years before, was reduced to four to years in the last census.
Now, we want to bring it down to two years. We hope to achieve a major release of tables by two years, 2013 or so.