I have no agenda at all. All I care about is what works. And don't think I'm defending planning in India. It is completely dysfunctional, one reason I only consult from outside now. As people have pointed out cars don't work everywhere. Under no circumstance can we abandon places like Pharganj. Yes we can plan new cities better with more space for traffic. And we are doing this but the gridlock continues because places like Pharganj are not being upgraded. The prohibitive cost of land however is telling us something about what we are doing. By the way I don't consider Paharganj a ghetto at all. I have been to Ghettos and Paharganj is not one. It is poor urban neighborhood but all I see is potential not a Ghetto.
The market is telling us land is a super duper premium component in India. And yes I know regulations distort this signal but even small towns are flashing the same signal. We will have to minimize our use of it. This is a market signal, ignore at your peril.
So far I have avoided technical discussion. Let me bring up a simple one. When a planner allocates say 40% of passenger traffic to cars there are certain things that must be done. Let me point out 2 in a highly simplified striped down version. Say we allocate a level of service of E, the absolute lowest without complete fail. Massaland is typically LOS C/D or better. At E clearance requirements around a car is ~1M with a buffer of 2.5m in city traffic. Volume to capacity V/C permitted is 22% meaning 78% of road is buffer, open, turn lanes, etc to prevent grid lock even though 78% is technically classified as grid lock. Pedestrian space requirement is roughly 6 sqft per person.
So lets do a quick calculation on how much traffic space we need for a 10 sqkm area of Paharganj or about 10x250 = 2500 acres. Assume density of 25,000 per sqkm, though Paharganj is actually 40,000 and higher. So I will assume half the population has already been dispatched elsewhere.
So total population in that area is 250,000. Since 40% must possess cars, say 100,000 cars in that area. So just for parking say a small compact car we need, 2+1mx2.5+1m = 10.5 sqm. Per car. For 100,000 cars we need 10.5x100,000 ~ 1 million sqm. ~ 1,000,000/4,000 = 250 acres. So 10% of land must be set aside for car parking or traffic will gridlock. Now keep in mind that typically you ned 2 to 2.5 times per car as you need to be able to drive somewhere to park the car and then drive back to park the car. But I'll ignore that. And multilevel car parks don't help because they displace habitation making densities even higher.
Let take two - traffic space needed for LOS E with V/C of 22/78. so 10.5x78/22= 37 sqm. So 37x100,000 = 3.7million sqm. So far we are up to 10%+37% = 47%. Or 47% of space in Paharganj will have to be demolished to accommodate cars for just 40% of the folks.
So lets take pedestrian needs, since people walk too. So 0.5x250,000 = 125,000 sqm. So a total of 40 acres! If we increase LOS to say pedestrian B we get 2.5x250,000=325,000 sqm. Or roughly 100, acres. About 4% of the land for LOS B, with parks, benches, streams, trees, fountains, etc.
With LOS B Public transport system. Typically 100,000 PPHPD, 2 metro lines on the diagonal could make sure everyone was within a km of a station. Cost is roughly 10 kmx2.5 Billion=Rs25 Billion. This would require 0 area if underground or say 1% if above.
This I think is doable.
You may not have said it but by calling Paharganj a ghetto and calling for demolition and car traffic that is the impact of your statement. You should not hide from it, nothing to be ashamed of as it is an arm of urban up gradation. Many many people advocate it. Undoubtedly some areas are beyond redemption and demolition will be required. Most buildings will in any case have to be demolished and rebuilt over 20-30 years.
Finally let me take up the picture. Those 20 cars occupy 10.5x20= 210 sqm of space. In a dense market like that with typically 3-4 persons per sqm that means 210x3 = 630 persons. Because of those 20 cars 630 persons were displaced. I would call that fat cat behavior, No.