Original Link :

Click to access SCIENCESCOPE_CAREERS_FEB%202017_low%20RES.pdf

Prof Sonali Das, a statistician who joined the CSIR, South Africa in 2007, says what she enjoys most about her career is the fact that her expertise can be applied to a wide variety of domains.

How did you choose an offbeat, unconventional and  unusual career such as statistics?

Statistics is one of the few subjects that allows you to collaborate with almost anybody who works with quantitative data. That is obviously almost everybody in the world of science, for example geologists, climate scientists, those who work in finance and economics, and even sports scientists. The more uncertainty about the data generating process, the more likely it is that a scientist will request the services of an expert statistician to help them with the interpretation.

What is your background?

Das was born in India and graduated with an MSc in statistics from the University of Calcutta in 1994. Thereafter, she joined the Indian Statistical Institute (ISI) as a computer programmer.

“I was surrounded by peers doingtheir PhDs, teaching me things and encouraging me to pursue new opportunities. I built close relationships with people who later moved to great academic positions abroad. I am still in contact with them for advice. This taught me that you learn from all of your experiences and it is better to be someone small surrounded by people who are more intelligent than you, than being a king amongst mediocre minds.”

In her free time, Das completed extra training in computer applications to better her technical programming skills and thenmoved to New Delhi to work in the government sector.

What did you do next?

“I soon realised that to climb the career ladder, I needed a PhD and this sent me back to my coresubject of statistics.” Das was accepted at the University of Connecticut in the USA where she completed her PhD in statistics in 2006, focusing on Bayesian statistics, which has also been one of her major research focus areas since she joined the CSIR. The Bayesian statistical approach allows a researcher to update their confidence or belief (in terms of probability) in the initial assumptions of their model with new data as it becomes available. In the game of heads and tails for example, this would mean that the result is not the number of times that heads will show up if the coin is tossed repeatedly, but rather a measure of confidence that a head will show up on the next toss.

What do you specialize in?

Das has worked on a variety of projects in different domains. These included doing risk analysis for climate change – helping researchers to quantify the carbon dioxide in the Southern Ocean with a higher degree of certainty. For this, she needed to understand the biology and chemistry of the carbon cycle. She also uses her modelling capabilities for sector-specific skills forecasting and to investigate how novel data modelling techniques could be applied to human movement analysis research that can help researchers to develop better treatment methods for people with disabilities and injuries.

“I greatly appreciate the fact that the CSIR provides me with opportunities to understand and contribute towards the science landscape with some of the best people in the field,” she says.

What skills are needed to be a statistician?

Das says statisticians have to excel in mathematics and be able to persevere and adapt. They should also have the willingness to make an effort to understand the various domains in which their input is needed. She says the field of statistics is becoming increasingly sophisticated, and one should be willing to learn the emerging areas
within statistics if one wants to be in a research career.

What are your qualifications?

Das has a BSc Honours (Statistics) from the Presidency College in India, an MSc (Statistics) from the University of Calcutta and a PhD (Statistics) from the University of Connecticut in the USA. Most South African universities offer degrees in statistics. Das advises that students do as many of the foundation courses as possible and that they should only embark on their PhD studies when they have a sound knowledge of the foundation of statistical theory and methods.