Data collection - Neighbourhoods in Mumbai

We are a property search engine and we need to enable our users in India to search for properties by neighbourhood (e.g. Juhu in Mumbai; Shivaji Nagar in Bangalore; Triplicane in Chennai). To do this, we need basic coordinates of all sub-divisions, neighbourhoods and other areas within Mumbai which our users might use to search our database. The simple basic test to determine which neighbourhoods we need is the question "would someone look for a property in that area?". We are not experts on local geography, but we have a rough list of the neighbourhoods we expect to have in our database eventually. For example, I would expect that most of the neighbourhoods mentioned on this Google Maps should be included: [login to view URL] To determine that the task is completed adequately, we will randomly take four or five neighbourhoods we would expect to have and we would check that the final deliverable indeed has them. Although we are currently posting this task only for Mumbai, we have the same requirements for other top cities in India. We'd be very happy to continue with you should everything works out fine in this test case - with five more cities in India to be mapped. To be specific, we need the following: * Bounding box for every neighbourhood in the City containing: minimum latitude, minimum longitude, maximum latitude, maximum longitude * Bounding box coordinates to be in decimal degrees. Here is an example: 12.97243,80.13752,13.09417,80.262482 . Note that we need this data resolution - rounding to "12.97, 80.14, 13.09,80.26" is not adequate * We do not need the entire city to be covered by neighbourhoods. We expect that there will be some gaps, but not the entire sectors missing. Basically, if there is an area where people do not look for property - we would expect to have it in the final deliverable. * We consider Mumbai to cover the following area - Mumbai District and Mumbai Suburban District: Mumbai District 'max_lat' => '19.05402', 'max_long' => '72.893707', 'min_lat' => '18.878559', 'min_long' => '72.790878' Mumbai Suburban District: 'max_lat' => '19.273821', 'max_long' => '72.978432', 'min_lat' => '18.989799', 'min_long' => '72.773331', This clearly covers some of the fringe areas. We can map neighbourhoods here more broadly - with closer resolution in the urban center. * Bounding boxes of neighbourhoods can overlap each other. In fact, we expect that they will almost certainly do so. Bounding boxes should cover the entire extent of the neighbourhood. This means that there will be a lot of overlap between the neighbourhoods. This is fine. * Bounding box of one neighbourhood can be entirely contained within a bounding box of another neighbourhood. For example, Karol Bagh neighbourhood in Delhi falls entirely within the neighbourhood of Old Delhi. * I am attaching the CSV file as a delivery template. We don't care how you collect the data, but the final output should be as specified in the attachment. Basic structure is simple - name, bounding box coordinates (use : as delimitor, coordinates go in minlong, minlat, maxlong, maxlat order) and vertical linking column should read ‘Mumbai' in all fields. The example attached still has some data from Chennai, but it should give you a better idea of what are we looking for. * We don't care how you go about data collection, but we find that online mapping applications such as [login to view URL], [login to view URL], or perhaps some of Google Mapping tools. * How do we know that the task is completed? Send us the final deliverable. We'll do a random test looking for the presence of several well-known neighbourhoods we know of in the city. Again, we believe that the neighbourhood shown on this Google Maps page are all required: [login to view URL]