Predicting the Present – Extended

Hal Varian’s Predicting the Present talks about how our Google search queries can be used to build a database of intentions. Search queries provide insights into people’s interests, intentions, and future actions.

An interested example cited in the article is that the most Google searches for the world “hangover” occur on Sundays; and that the most searches for “vodka” peak every Saturday (one day before the “hangover” peak). The exception to this regular weekly pattern occurs once a year, on New Year’s Eve.

Likewise, users tend to search for influenza-like symptoms when they have the flu (see: Detecting influenza epidemics using search engine query data). And if a user searches for Nvidia GTX670, chances are they are looking to buy a particular video card (tip: keywords such as particular product model numbers tend to have high click-through-rates on Google AdWords).

Interestingly, Google’s “database of intentions” can go much deeper than that…

Our Gmail accounts may reveal information such as our flight itineraries and shipping packages (along with carrier tracking numbers) – giving Google the data to predict where/when we are going to fly or when we are expecting a package. This data can translate in to features – such as a prominently displaying a flight status tracker or a shipping tracker on a Google dashboard (such as your Android phone’s “Google Now” cards).

Our Gmail and Android Phones reveal which contacts we contact the most. This provides great data on who our closest friends, family, and colleagues are (which can be easily differentiated by analyzing the times we communicate, and the body content and CC: data in our e-mail messages). I believe we are more likely to contact our colleagues during working-hours and our friends/family during our off-hours. And we are prone to CC’ing colleagues who have some relationship with each other (for example, when I send out an e-mail to my development team’s product manager, the contacts I would cc: most often would probably include my team’s UX designer, and a couple other software engineers). This contact data can be used to recommend who I should add to my Google+ circles, which can then be ultimately used to help personalize my search results and advertisements.

Our Google Calendars may contain a detailed schedule of when and what classes or work I need to complete, or perhaps what activities I intent to involve myself in. If Google knows I will be going on a date on Saturday night, it would make a lot of sense to advertise places to go Friday night or Saturday afternoon.

Google crowd-sources our Android GPS data in real-time to provide traffic data to Google Maps (example: what is the average speed on I-85 right now). The location based services or GPS data on our Android phones reveal what times you spend at work or home. The data can also be used to predict when and what stores you like to shop at regularly for advertisements, or even used to catch a cheating spouse.

A Minority Report styled future may be becoming closer and closer to reality. 

I believe it is important for us to think about the amount of information Google (or any other popular website, such as Facebook, Amazon, Reddit, Bing, etc.) is collecting on us and how it is used, because I believe that data can be incredibly valuable and powerful. Big-data enables companies (such as Google) to predict our intentions and actions before we act.