An art or a science?

A Full Bengali Meal

Is cooking an art or a science? The debate goes on. Some say it is the ideal meeting point of art and science. The way any recipe is presented is more akin to science – ingredients required, the quantities, the sequence of activities, the temperature to be maintained – sounds almost like a chemistry experiment!!!

But once you execute the process it starts resembling an art where you have to use most of your senses. You need to see if your creation looks beautiful, hear the crackle of spices in hot oil, enjoy the lovely smell (better to call it aroma), taste and adjust the seasoning and feel the texture. An activity where you use all your five senses to such an extent has to be an art right!!!

For beginners, cooking is more like a science where you depend on the measures and timings and try to get the final dish as per the copy book. However, as you progress things start becoming andaaz se – approximately, as per your gut, intorno a !!! If you have noticed, all your grand mom’s recipes would be andaaz se – more an art than a science.

So where does analytics fit into this? Well analytics – or for that matter any research be it qualitative or quantitative is finally interpretation of findings. A friend of mine who is a senior executive in a leading quantitative research firm told me that the findings keep changing (improving) as the data moves up the value chain. Several times its requires his intervention to find insights which will be meaningful for the client. A qualitative research analyst once mentioned that the same data would give him a different insight if seen the first thing in the morning as compared to the last thing in the evening!!! This sounds more like an art than a science!!

Technology tries to reduce the subjectivity in interpretation of data. Algorithms are designed to drive the thought process toward a similar if not same interpretation.

Use of science in cooking has come a long way – on one end of the spectrum is molecular gastronomy and at the other end is industrial grade production of fast food. Similarly in analytics, especially that of unstructured data, Natural Language Processing (NLP) and other such artificial intelligence(AI) tools have brought about significant ‘structuring’ which enhances efficiency (by reducing time take for analysis) and accuracy (by removing subjectivity in interpretation).  But NLP and AI still have some way to go.

An ontology represents knowledge as a set of concepts within a domain and the relationships between those concepts.

My take on this is technology in analytics is here to stay (like the food processor, refrigerator, automatic oven etc in your kitchen). As time progresses technology will become more intuitive and imperative (and you will add the liquid nitrogen container, heated bath, rotary evaporator etc to your kitchen!!!). But as they say in Hindi – Ghar ke khane ka maaza hi kuch aur hai (roughly translates to ‘There is a different kind of joy in having home cooked food’)

Ameet Das

The author is the Co Founder of Vondasoft Technologies which is a Mumbai based technology start-up. Vondasoft has developed an algorithm for unstructured data analysis and specializes on Social Listening and Analytics.

Picture credits:

"Full meal" by Nandinissaha - Own work. Licensed under Creative Commons Attribution-Share Alike 3.0 via Wikimedia Commons

"AlineaDessert" by Brancron (talk) (Uploads) - Own work. Via Wikipedia

"GFO taxonomy tree"  Original uploader was Leechuck at en.wikipedia -  Licensed under Creative Commons Attribution-Share Alike 3.0 via Wikimedia Commons 

Is Less Really More?

A marked difference between the Indian cooking style and the Western way of cooking (which would include say Italian, French, Mediterranean, Continental to count a few) is the number of spices/elements used in a recipe.  In a recent episode on Fox Life – David Rocco’s Dolce India, there was a face-off between David and Chef Rajdeep of ITC Hotels – friendly of course. They both decided to make meat balls. The fact that Chef Rajdeep’s meat balls are better known as Galouti Kebabs is another issue !!! While the Indian Chef used about 20+ spices his Italian counterpart used only about five.

 

This gives rise to a favorite debate amongst foodies – who has a more evolved palate? The Indian foodies like to believe that their sense of taste is more evolved – their taste buds can pick out several distinct flavors in one dish. The Europeans on the other hand feel that a dish should have one distinct flavor which is its signature or identity. At best there can be a second flavor which comes as an after taste. Who is right and who is wrong – well your point of view is as good as mine!!!

In case you are wondering why I am talking about food on this forum – for one it is a passion for me;  but more importantly we come across a similar issue with analytics of unstructured data. Unstructured data is by definition – well, unstructured. Therefore, any reports or insights you try to get from this data is not defined by strict rules. The question which the analyst faces on a regular basis – should I make just one or at best two inferences from this myriad of information (the European palate) or do I look at all aspects that this data provides to get multitude inference which may or may not be actionable (the Indian taste).

Similar to the cooking styles, we believe this does not have a right or a wrong answer. Ideally, the defined objective of the exercise should give us the direction to follow. For example – if it is a hypothesis that is being tested then the approach has to be very focused without any scope for distractions from noise. However, if the objective is to discover unknown unknowns then we have to look at all possible inferences which the data throws up. You never know which discovery may be the most important one!!

Afterthought – Culinary tastes and cooking techniques are not a reflection of anyone’s approach towards analytics…the analogy has been drawn to make the reading more interesting!!

The author is the Co Founder of Vondasoft Technologies which is a Mumbai based technology start-up. Vondasoft has developed an algorithm for unstructured data analysis and specializes on Social Listening and Analytics.