I turned to the database and started querying all the data about the trips ending on that cluster.
To my amazement, I found that drivers would stop at the middle of nowhere for thirty to forty minutes, and the distribution of these stop durations had a small variance.
Just what on Earth were they doing there?Coin DropWell, you probably figured it out by now.
These drivers were attending to what I came to call “roadside entertainment” activities.
When the coin finally dropped on me, I was shocked.
Not because I was witnessing the data trails of one of the oldest professions in the world (some say spying is the oldest), but because with a little extra effort I could associate a driver name to each of those data points.
With a relatively simple data science exercise and with the right (or wrong) data, I could have had access to the private lives of others.
Having this type of unauthorized access is wrong in any way I can think of it.
I quickly dropped that ball and moved on to build the trip graph.
The trip graph became the data foundation for many exciting data science products, from trip anomaly detection to destination prediction.
However, I never forgot the sense of awe and fear that I experienced after realizing the power of a simple data analysis.
I felt awe at the power that a simple data analysis can unleash, and fear that such powers could be used on me, or against me.
I am now a firm believer that the data science community must actively promote privacy best practices, and must also keep vigilant to prevent unintentional uses of analytics in revealing what must be kept private: our personal lives.
No matter what.
.. More details