Given the traffic snarls and developing infrastructure, I would love to know how many months or years it ll take them to get this done. My other concern is whether they ll be able to cope up with India's changing infrastructure. New bridges, buildings, housing complexes come up every year, traffic patterns change to accomodate the same and most big cities have a metro under construction. Changing maps is one thing, but re-mapping street view is quite another.
Interestingly, I was discussing this with some people while in India two years ago.
It turns out that a lot of houses don't really have addresses. A lot of alleys don't have street names. A lot of buildings just get built, and there's not always an official address for them. So it would definitely be interesting, yes.
Google actually came up with landmark-based directions specifically for India( http://googleblog.blogspot.com/2009/12/go-thataway-google-ma... ), but if you do not have an official address I guess one would have to settle for a nearby intersection or business.
The good thing for a project like this in India is that labor is relatively cheap. Hiring semi-skilled technicians / drivers to drive the street view camera van will probably not cost much. Back of the envelope calculation:
~3M Km of roads (urban and rural)
- Assuming avg speed of 30KM/hr, that is 100,000 man hours of work just to record the data.
- Double the estimate to accomodate 2 person teams (driver and technician). 200,000 man hours.
- 50 Rs/hour on average would probably be decent pay for a driver/technician that's semiskilled. (Rs 12000/month with an 8 hour workday). 50 Rs > $1. So, let's say $1 per hour.
- Total cost for recording ALL urban and rural roads excluding state and national highways: $200,000
That's probably around the annual salary of one Sr. engineer at Google in Mountain View.
They could afford to do it multiple times a year.
This doesn't consider the cost of processing this data but Google being Google, I am sure they have automated tools that convert these recordings into Street view quite cheaply. Computing power is their cheapest resource.
Improving the estimate:
- There's also the cost of equipment. It would take 1000 teams 12 working days (~100 hours) each to record data. If you have a little more than a month you only need 333 teams, and so on. Also I bet they are only doing cities first, so they could probably do with 100 teams at a time. Let's say the equipment is USD $2K per team. That's $200,000 in equipment. I am guessing they can do it for much cheaper since they are likely reusing equipment across countries.
- There's the cost of renting the auto-rickshaws. Looks like the 'retail' rate for an auto-rickshaw trip is about Rs 12 per KM (all inclusive) for short trips. Since we are paying the driver separately, 12 per KM may be excessive for the rickshaw + gas, but let's use it anyway. That's Another Rs 12,00,000 = ~$24,000
Updated total cost: $424,000
Imagine if you were a startup in India circa 2007 when street view launched in the U.S. If you had the technical chops to build tools for street view, it would have been a no-brainer to invest this kind of money and get the data. Google, Microsoft or Yahoo would have happily paid 10x to acquire the startup.
I haven't been here that long, but from what I've seen, Google doesn't even have up to date maps in India. I went jogging a few days ago, and ran into what google maps described as a dead end. The road didn't stop - it just continued through some "undocumented housing" in dirt road form.
Those "tri-cycles" that the articles mentions are called "autos" in India. They are usually used as taxis. It's interesting how Google chose to use them rather than a car! A good example of localization!
Now those babies will get you through all the back-alleys and streets in no time. The only problem would be mounting their camera on it. I think they ve solved harder problems.
They are not using autos! It's a "trike" that's manually driven by the rider. If you notice carefully in the original article, there are no cameras above the auto.
I love Street View and cannot believe I'm living in a time and place in history that makes it possible to see such a massive amount of culture and everyday life from virtually anywhere in the world.
One question I've had since Street View was first released is how does this make business sense for Google? I can see how it relates to Google's stated mission statement ("to organize the world's information and make it universally accessible and useful"). But how can Google justify spending vast amounts of cash on what seems to be a glorified pet project? Is it simply a desire to build the best mapping service, thereby attracting visitors and eventually generating advertising revenue?
btw, I'm not trolling or flaming -- it's been a genuine curiosity of mine for years now.
I don't have an answer to your actual question about Google's motivations here, but I do think Street View is valuable, and what's genuinely valuable for users, eventually leads to money for the provider.
However, I am genuinely curious why you had to qualify your very reasonable question with a "I am not trolling or flaming". I have noticed many other commenters do this recently, and don't understand the need for qualifying yourself if you have a genuine question.
Yes, I agree that it's valuable and will eventually lead to revenue. Perhaps the "mystery" stops right there.
Sometimes it can be difficult to clearly communicate in short forum-style posts. I don't always qualify my posts, but in this case, I didn't want HN readers to think my posts boiled down to ripping on this Google Street View announcement. I should let my posts speak for themselves and work on being more concise. ;)
I feel like there's been an increase in this type of behavior since points were removed from individual posts.
If you think about all the hype around "local," Street View is a massive and near-impossible to duplicate asset in that battle. With some good image recognition of signs, people, etc. Google could map out specific business locations, exact address locations, which combined with real-time location information from mobile Maps users can make Google Maps basically undefeatable (and it has a solid business model).
The attack plan is simple: if you build it, they will come.
Let's say I wanted to build a local business service. I could do all of the searching myself, or I could pay some money to google to leverage their platform. There is instant recognition of the google service, and it helps those startups decouple the difficulties of information acquisition from the business model.
In general, google, facebook, apple all do the same thing: maintain a platform upon which others build stuff.
This is all an (informed) guess, but here's how I imagine it happened:
2005-ish MapQuest, Google Maps, etc. were all using Navteq and TeleAtlas as their data providers. Google was probably paying millions of dollars per year for licensing Navteq and TeleAtlas' entire data sets. They were probably dumping hundreds of expensive engineer-hours into merging these datasets, applying updates every quarter, dealing with customer feedback, etc.
In 2006 or so they probably did some napkin-math and saw that a) owning their own data would let them do whatever the heck they wanted with their maps (I bet Navteq wanted much more money for the ability to sell ads on top of their data) and b) over time, it would be cheaper to spend a few million shipping cars and paying drivers to collect their own data than handle the increasingly-unreliable commercial data vendors (remember that at the time the commercial vendors didn't have any competition and so weren't doing much more than driving new roads every few months). Also, for the most part geo-data is a one-time cost. The majority of the world's road network stays the same and further updates can be applied using customer feedback (MapMaker) or aerial imagery. Oh, they also had this handset project coming out called "Android" that needed access to a non-GPS method of determining location, so they strapped some wifi cards onto the cars.
I don't think there's any question that the millions of dollars they spent to collect and index the physical data of the world was worth it. It improved every single one of their core products: Maps, Earth, Ads, Local, Places, etc. all sprouted from this data that they collected. Also, the huge leap they took is slowly being replicated by the commercial vendors (Navteq started driving their version of StreetView a few months after Google announced their first city).
This is very interesting, and I LOVE that Google is doing it. I use google maps to make sure that a rickshaw/taxi driver isn't ripping me off by going the wrong/long way. It has come in handy for me several times in India.
The real trouble in India is addresses. When I lived in India my official address was
3rd Floor Pran Vijay / Near H K House Lane / Opposite Bata Showroom / Income Tax Colony / Ashram Road / Navrangpura Ahmedabad-380009
How the hell will Gmaps figure that out? Will we just start using lat/long coordinates?
Lot of companies in India both public and private, including the govt. have subsidized housing adjacent to the office buildings. They're often called colonies. A lot of the times, these subsidized housing options are open to the public as well, maybe at higher rates. This might be one instance relating to the IT (Income Tax) Dept.
yes mapping India will prove to be quite the challenge - there are a lot of roads and alleys that are undocumented. Interestingly, the street my old house was on never shows up in Google, shows up on Yahoo (but is unnamed) & shows up named on Bing. I've noticed such discrepancies with other small streets and alleys as well
Which leads me to wonder - how do companies gather map data (as in street names, building names etc) ? Do they do it manually or parse existing maps?
27 comments
[ 2.7 ms ] story [ 72.8 ms ] threadIt turns out that a lot of houses don't really have addresses. A lot of alleys don't have street names. A lot of buildings just get built, and there's not always an official address for them. So it would definitely be interesting, yes.
~3M Km of roads (urban and rural)
- Assuming avg speed of 30KM/hr, that is 100,000 man hours of work just to record the data.
- Double the estimate to accomodate 2 person teams (driver and technician). 200,000 man hours.
- 50 Rs/hour on average would probably be decent pay for a driver/technician that's semiskilled. (Rs 12000/month with an 8 hour workday). 50 Rs > $1. So, let's say $1 per hour.
- Total cost for recording ALL urban and rural roads excluding state and national highways: $200,000
That's probably around the annual salary of one Sr. engineer at Google in Mountain View.
They could afford to do it multiple times a year.
This doesn't consider the cost of processing this data but Google being Google, I am sure they have automated tools that convert these recordings into Street view quite cheaply. Computing power is their cheapest resource.
Improving the estimate:
- There's also the cost of equipment. It would take 1000 teams 12 working days (~100 hours) each to record data. If you have a little more than a month you only need 333 teams, and so on. Also I bet they are only doing cities first, so they could probably do with 100 teams at a time. Let's say the equipment is USD $2K per team. That's $200,000 in equipment. I am guessing they can do it for much cheaper since they are likely reusing equipment across countries.
- There's the cost of renting the auto-rickshaws. Looks like the 'retail' rate for an auto-rickshaw trip is about Rs 12 per KM (all inclusive) for short trips. Since we are paying the driver separately, 12 per KM may be excessive for the rickshaw + gas, but let's use it anyway. That's Another Rs 12,00,000 = ~$24,000
Updated total cost: $424,000
Imagine if you were a startup in India circa 2007 when street view launched in the U.S. If you had the technical chops to build tools for street view, it would have been a no-brainer to invest this kind of money and get the data. Google, Microsoft or Yahoo would have happily paid 10x to acquire the startup.
Google has their work cut out for them.
Now those babies will get you through all the back-alleys and streets in no time. The only problem would be mounting their camera on it. I think they ve solved harder problems.
http://googleindia.blogspot.com/2011/05/google-street-view-s...
http://gadgets.boingboing.net/2009/05/18/google-street-view-...
http://www.youtube.com/watch?v=Hr-4Aln1Il8&playnext=1...
One question I've had since Street View was first released is how does this make business sense for Google? I can see how it relates to Google's stated mission statement ("to organize the world's information and make it universally accessible and useful"). But how can Google justify spending vast amounts of cash on what seems to be a glorified pet project? Is it simply a desire to build the best mapping service, thereby attracting visitors and eventually generating advertising revenue?
btw, I'm not trolling or flaming -- it's been a genuine curiosity of mine for years now.
However, I am genuinely curious why you had to qualify your very reasonable question with a "I am not trolling or flaming". I have noticed many other commenters do this recently, and don't understand the need for qualifying yourself if you have a genuine question.
Btw, I am not trolling or... ;)
Sometimes it can be difficult to clearly communicate in short forum-style posts. I don't always qualify my posts, but in this case, I didn't want HN readers to think my posts boiled down to ripping on this Google Street View announcement. I should let my posts speak for themselves and work on being more concise. ;)
I feel like there's been an increase in this type of behavior since points were removed from individual posts.
Google can afford to play the "long game."
Let's say I wanted to build a local business service. I could do all of the searching myself, or I could pay some money to google to leverage their platform. There is instant recognition of the google service, and it helps those startups decouple the difficulties of information acquisition from the business model.
In general, google, facebook, apple all do the same thing: maintain a platform upon which others build stuff.
2005-ish MapQuest, Google Maps, etc. were all using Navteq and TeleAtlas as their data providers. Google was probably paying millions of dollars per year for licensing Navteq and TeleAtlas' entire data sets. They were probably dumping hundreds of expensive engineer-hours into merging these datasets, applying updates every quarter, dealing with customer feedback, etc.
In 2006 or so they probably did some napkin-math and saw that a) owning their own data would let them do whatever the heck they wanted with their maps (I bet Navteq wanted much more money for the ability to sell ads on top of their data) and b) over time, it would be cheaper to spend a few million shipping cars and paying drivers to collect their own data than handle the increasingly-unreliable commercial data vendors (remember that at the time the commercial vendors didn't have any competition and so weren't doing much more than driving new roads every few months). Also, for the most part geo-data is a one-time cost. The majority of the world's road network stays the same and further updates can be applied using customer feedback (MapMaker) or aerial imagery. Oh, they also had this handset project coming out called "Android" that needed access to a non-GPS method of determining location, so they strapped some wifi cards onto the cars.
I don't think there's any question that the millions of dollars they spent to collect and index the physical data of the world was worth it. It improved every single one of their core products: Maps, Earth, Ads, Local, Places, etc. all sprouted from this data that they collected. Also, the huge leap they took is slowly being replicated by the commercial vendors (Navteq started driving their version of StreetView a few months after Google announced their first city).
The real trouble in India is addresses. When I lived in India my official address was 3rd Floor Pran Vijay / Near H K House Lane / Opposite Bata Showroom / Income Tax Colony / Ashram Road / Navrangpura Ahmedabad-380009
How the hell will Gmaps figure that out? Will we just start using lat/long coordinates?
Lot of companies in India both public and private, including the govt. have subsidized housing adjacent to the office buildings. They're often called colonies. A lot of the times, these subsidized housing options are open to the public as well, maybe at higher rates. This might be one instance relating to the IT (Income Tax) Dept.
Which leads me to wonder - how do companies gather map data (as in street names, building names etc) ? Do they do it manually or parse existing maps?