Just a random thought.
I walked into the gym this morning and swiped my membership card. There's a data cluster with my name, a picture of my face, all the information I submitted with my gym membership, and I just added my location on that date and time.
Later on, I use my credit card at a shopping center, where again the date, time, and location are recorded along with my purchase. It's the third time this month I've purchased hummus, so that preference is added to my data.
Since my gym routine has been recorded as habitual, that point of data is scored on a certainty scale of 1-100 as a 94, easily scored even higher when the facial-recognition algorithm of the gym's security camera matches the one on my gym ID card. However, the store's camera wasn't tied into the database, but since it was a grocery store and I hadn't been to a grocery store in ten days, that point of data was scored with a certainty of 72.
Now there's two points of location data for me, one with near absolute certainty and another with somewhat certainty. Given that, a certain radius of travel can be drawn for me, with its own certainty score.
My name and data are also associated with my car's VPN and license plate. While driving to the store, I stopped at a red light, where a camera recorded my license plate, along with my location, the date, and time. Now another point of location data has been recorded, with a certainty score of about 83. (Was his partner driving the car? Was it stolen? Did he loan it to someone?)
Using three points of location data with three different times, my locations between those times is narrowed even further, with assumed points along the route given certainty scores.
Brian has a cell phone conversation with a mutual friend of ours, during which he says the sentence, "Adam went to the store to pick up some celery, and guess who he ran into? Yes, it was Bob, so they had lunch."
The conversation is digitized, analyzed. I am a known associate of Brian's so the data relationship has a high certainty score. "Went" and "store" are noted but have no time or date associated with them except for the date they were said. "Celery" is picked up, and so is "Bob" and "lunch."
That data is clustered together into its own cluster, each piece of which with its own certainty score.
Bob and Adam have lunch. Bob picked up the check, and the number of diners is noted. The date, time, location, and food ordered are all noted. This data, combined with the data from Brian's phone call, combined with Adam's travel radius for the day, all has its own certainty scores and adds another point to Adam's travel radius, narrowing his path down even further.
Like a giant jigsaw puzzle -- more a fluid mosaic, the model used for evolution paths -- our data could easily be centralized and ever-populated with facts about us, facts with a wide range of certainty scores -- scores that are constantly changing based on the new data collected, scores that increase or decrease based on their relationship with other data and the certainty scores of that other data.
Just a random thought.