The tool proved so popular it exceeded the API limits on Mr Khanna’s Mapbox account.
So Mr Khanna has shared the code for the map on a Github page so people with coding knowledge can get it working.
Testing his map on a group of distant Facebook friends who posted on chat at least once a day, Mr Khanna found that by looking at clusters of messages, he could see where one of them lived – even down to the location of his dorm room.
By looking at two weeks of chat data, he managed to work out the Stanford student’s weekly schedule, which means he could predict which building he would leave on campus at a certain time – information that could prove potentially dangerous in the wrong hands.
‘In fact I found that I could infer a schedule for almost everyone in this chat as well as the other active chats I am in,’ he wrote.
Mr Khanna found that he could even do this for people who he is not friends with on Facebook and could track the locations accurately from messages sent within a large group he was a member of, in order to organise poker games.
After chatting frequently to one friend, he even discovered that using the map, he could track his hourly movements so that at the end of the day, the map matched the location history collected by his phone.
