Are food and political preferences correlated?

In early April 2016, Foursquare predicted the Mexican food restaurant chain Chipotle had suffered a 30 per cent decline in sales between January-march 2016. Chipotle’s results, released a few days later, showe d sale s were down 2 8.4 p er cent. Foursquare had made similar, eerily accurate estimates earlier—it had correctly projected the number of iphones sold on a weekend for example.

Foursquare claims to be the “mo st trusted i ndep endent data location platform” and it provides popular online city guides. Its data is based on app-users “checking in” at specific locations, recommendations of places to visit, etc. Its insight into Chip otle’s sale s, and Apple’s we ekend s ale s, was based on analysis of this data.

This is one example of how data can be mined to offer unusual, actionable insights. Another example arose from a recent survey conducted by The New York Times and Siena College in the US state of Iowa. When 584 Democrat voters were asked about their culinary preferences, it transpired that those who liked Indian food were more likely to vote for Bernie Sanders, while those who did not eat Indian food were more likely to vote for Joe Biden in the Democratic Primaries. Just as interestingly, registered Republicans who had been to Europe, Australia, Canada, or M exico, or ate at Indian restaurants, were less enamoured of Donald Trump, than Republicans who had not travelled outside the US, or eaten Indian food.

Food preferences are more obviously political in India. The Bharatiya Janata Par ty (BJP) has associated its brand with veneration of the cow, and more broadly with vegetarianism. No t only do BJP governments tr y to ban beef anywhere they have leverage; Bjp-ruled states have banned sale of meat during Jain festivals and stopped serving eggs to under-nourished children in s cho ol midday meals. The lynching of Muslims on the grounds that they are b eefeaters is also quite a popular pastime in Bjp-ruled states.

Vegetarianism is generally associated with upper castes. But rising incomes over the last two decades has translated into higher meat and egg consumption. Household surveys and s elf- declarations in the Census 2011 suggest that meateating is far more prevalent than t he BJP ’s i de olo gue s would have us believe.

The 2011 Census data indicates that only about 30 per cent of Indians above 15 declare themselves to be vegetarian – and this may b e an over-statement made due to social pressures. The number varies a lot from state to state, and caste to caste.

There may be a correlation between the BJP vote -share and vegetarianism. India has five states with around half the population, or more, claiming to be vegetarian. These are Rajasthan, Haryana, Gujarat, Madhya Pradesh ( MP) and P unjab. Gujarat and Har yana are currently BJPruled, while Rajasthan and MP had BJP governments for years. Even Punjab has endured a BJPAkali Dal coalition. All five of the s e state s have stable BJP vote share, which seem to map to the number of vegetarians.

India also has eight states, where less than 10 per cent of the population claims to be vegetarian. Out of Telangana, West Bengal, Andhra Pradesh, Tamil Nadu, Kerala, O disha, Jharkhand and Bihar, only Bihar has a BJP presence, as a coalition partner. The only Southern state where the BJP has a large vote share is Karnataka, where vegetarians comprise 21 per cent of the population.

We need more granular, hyper-locational data to see if there is a strong overlap in votes and diet, of course. Such data could perhaps be available with food delivery firms like Swig gy and Z omato. It would be interesting to analyse this for Delhi and other metros, where there maybe pin codes and, therefore, constituencies that are more, or less, vegetarian.

Would vote share map out the same way? Does more meat consumption translate into fewer vote s for the BJP, and vice -versa; does lower meat consumption lead to more votes? It’ s an intriguing thought. Given a correlation between rising incomes and more meat consumption, acceptance of this hypothesis could also lead to Chanakyastyle justifications for tanking the economy.

We need more granular, hyper-locational data to see if there is a strong overlap in votes and diet. Such data could perhaps be available with food delivery firms like Swiggy and Zomato. It would be interesting to analyse this for Delhi and other metros, where there may be pin codes and, therefore, constituencies that are more, or less, vegetarian