Luckily, the server farms in Oxford remained unaffected by the huge electricity blackout rocking North India for the last 48 hours. Ever since my post on Muslim names, they stoically crawled through Lucknow's current electoral rolls to guess the religious community each voter belongs to. They identified 98% of all voters at the pace of roughly half a million names a day (for comments on the accuracy of this matching exercise, see below). And they thus enabled me to generate a draft map of Muslim life in Lucknow by caculating population shares (voter population shares, to be precise) on polling booth level and linking them to polling station locality (drawn from the National Informatics Centre).

Apart from the river (in blue), my own home (red dot), and the MODIS built-up area polygon already featured in earlier maps, the following visualization shows "Muslim name" density (background shade), Haj pilgrims (green circles; drawn from the 2012 Qurrah), and major Islamic institutions in Lucknow. The map is interactive: do zoom in for increasing detail, and click on the mosques to get to the respective institution's website (there is also a larger version):

The map provides some expected and a number of unexpected insights. Voters with Muslim names tend to live in the old city - and not across the river in new Lucknow, as expected. The same is true for Haj pilgrims, and also for key Islamic institutions (the only two across the river being the reformist Nadwa and the Barelvi Warzia seminaries). More surprisingly, however, residential patterns are more segregated in new Lucknow than in the old city. While this is a general trend in midle-class India, I did not expect it to be that pronounced in Lucknow (which has not seen major communal tensions so far). While old Lucknow thus shows up in many shades of grey (no pun intended), new Lucknow is black-and-white (consider, for instance, the area just north of Warzia in Gomti Nagar - a ghetto of 70% Muslims surrounded by neighbourhoods with barely 5% Muslim population). I was also kind of surprised to see the majority of Haj pilgrims coming from the old city, too. True: more Muslims live there to start with - but then, they also tend to be poorer. Or so I thought - till people told me that a) there is quite some money in old Lucknow, though less visible than in the posher neighbourhoods across the river - and that b) Haj pilgrims need not necessarily stem from among the well-off. Many artisans (who tend to live and work in the old city) often save money for their whole life to make it to Mekka once...

These are, of course, just some very initial observations. Much more could be said, for instance, about the significance (or not) of certain Islamic institutions for certain sections of Lucknowi Muslims, or about the links (as well as tensions) among them. I will hold on to this for a while - partly because I have a number of interesting interviews about this map lined up. But feel free to give your own five cents in the comment box below!

One last point is important to make, however: this map quite deliberately includes not one, but three ways of looking at Muslim life: density of "Muslim names", location of Haj pilgrims, and major Islamic institutions. Partly, this multi-dimensionality is to counter the criticism that my name-to-community algorithm essentializes and objectifies Muslimness in a quite colonial way (a criticism raised by some colleagues, for which I am thankful even if I naturally disagree). Indeed: if one took the algorithm in isolation, and would deduce anything else than "person X seems to have a Muslim-sounding name" from its findings, this criticism would be quite appropriate. But I am not presuming any hierarchy among the various indicators here, nor do I want to stress names too much (even though some of my informants here emphasize them a lot).

Does one's being Muslim lie in one's name (and, by extension, lineage)? Or does it depend on the observance of key Islamic commandments (such as once taking the pilgrimage to Mekka if one can afford to do so)? Or is Muslimness nested within certain Islamic institutions, the clergy in particular? All three options are frequently mentioned in my conversations in Lucknow: some feel Muslim most when called by their name, some distinguish proper Muslims from less proper ones by ritualism, and others still defined their being Muslim primarily through their allegiance to Islamic institutions (be they Barelvi Pirs or Shia clerics, or even just the madrassah from which one graduated while still young). Most people, of course, draw on all three (and many more) sources to define their Muslimness. The multi-dimensionality of this map thus attempts to convey some of the multi-dimensionality of lived Muslim lifes.

Finally, some more technical caveats and explanations are in order. A number of remarks on the similarity (or not) between electoral rolls and the test corpus used to measure my algorithm's (surprisingly high) specificity and selectivity can be found my recent comment on the algorithm post. In addition to what I wrote there, electoral rolls obviously only include adults, under-represent women, and can be fudged - but then, it is unclear if this would be true to different extent for Muslims and non-Muslims, and while fudging the rolls is an undeniable practice, they are also most closely watched by various competing political parties, and thus the relatively best data source available.

The map presented in this post is thus far from an ideal representation of Muslim life in Lucknow; it is one among many ways of grasping the spatial distribution of "being Muslim". But it is, I believe, a rather compelling one. What do you think?

I would like to acknowledge the use of the Oxford Supercomputing Centre (OSC) in carrying out this work.