How is our data used?

It’s no secret that the entire world is worried about the privacy of citizens. There’re concerns that modern companies know way too much information about us. This post is my thought experiment where I try to explain how I think our data will be used by Google to deduce information(As mentioned, this is a thought experiment. I'm not sure if Google does this or not)

  Let’s start with a simple use case scenario. I want to setup a Kirana store and I approached Google to help me. How can Google help me? To begin with, it can tell me average cost to lease a store in different parts of city. This is child’s play. It can tell me the potential of a given area. For example, let’s assume if everyone in the colony is using google pay exclusively to make payments. Google knows the monthly expenditure of people in the colony. Furthermore, Google made it mandatory to turn on device location before making payment. So, Google can also tell me the revenue of each shop present in the colony. What else can I know from Google? It’s a common observation that people prefer to shop with someone who shares a similar background. For example, a Telugu person might prefer shopping with a Telugu shopkeeper. Google can deduce these patterns and tell me. It knows that I speak Telugu from my Youtube & browser search history. Even if I’m cautious and always use private browsing, it can deduce the IP address associated with my phone and correlate it with my searches. Similarly, it can deduce the mother tongue of the shopkeeper. From this information, it’s trivial to make the above connection.

  Is that all? No. Google can help me manage my inventory. For example, people from the colony might be going to a rice mart 5 KM away from the colony to purchase rice every month. It can be inferred that, no one is selling rice at cheap price in the colony. So, Google can suggest me to stock up rice and sell it at a particular price to make profits. Similarly, it can tell me exactly what items should I stock up on, what items are not necessary. 

  This is how browser history and location can be used to glean information from us. Now, let’s go a bit deeper. Can Google know who is dating whom? If I’ve to make a guess, I’d say yes. If I’ve the data that Google has about people, how would I figure this out? Let’s take a hypothetical case where I’ve to figure out who a person named Vijay is seeing. First thing I’d do is to guess the gender of person. That’s really easy. Vijay would’ve searched for ‘best razor to buy’. Or, he’d have searched for ‘Should guys remove their shirt before entering sanctum in Kerala’. It’s also trivial to know his sexual orientation from browser history. Let’s say that, he’s a straight guy. With these two data points, I’ve cut the search space for the potential girl friend. How do I proceed now? I will make a list of IP addresses of girls he has hung out a lot with. I also record the locations where he hung out with these girls. That cuts down the search space even further. I can eliminate most of his female colleagues by using the fact that he will be within their proximity mostly in office premises and not in restaurants and malls. Okay, a person might hang out a lot with siblings too. How do you distinguish a girlfriend from sibling? That’s easy. If Vijay is staying with his sister, both share the same home. They might have similar usernames. They might be located at same place on the day of Rakhi. There’re huge number of ways to deduce just based on data if the girl is his sister or not. That leaves us with the final question. How do we distinguish girlfriend from a girl friend. Again, we can deduce this from locations of people on valentines day, purchase history of the person before meeting the girl. We can also deduce this based on the kind of locations they meet.

  I personally think, any question about us can be answered comfortably with the kind of data Google/FB/Amazon has about us. You might be able to cheat yourself but you can't cheat Google. Beware!