TIDE (Tiled Device Estimates) provides de-identified counts of location requests and unique devices within an approximately 100x100 meter (hectare) tile for every hour of a week. The data is categorized by source type (IoT, App-based, etc) and localized into every timezone.
Resolving these locations generates a massive data set that describes aggregate mobile device behavior with highly precise latitude, longitude and timing information. This dataset allows for robust analysis applicable to facility siting, network buildout, public safety, crowd estimation, asset vulnerability assessments, and other applications around the world.
All location information is gathered anonymously through the normal operation of the Skyhook Precision Location Service or acquired through partnerships with app developers and aggregators. Skyhook anonymizes location requests from their original devices by removal of all persistent identifiers. Aggregating device behavior by hour adds additional k-anonymity within any given tile. This process ensures privacy at the device level yet provides an analytically potent view of activity levels in virtually every location where people are carrying smartphones and other WiFi-enabled devices.
Crowd detection, origin and destination mapping to high activity areas.
Example: How does crowd density correlate to known incidents of violent crimes?
Activity measurement for targeting fiber build-outs, cell antenna expansion and optimization of communication and data networks.
Example: What are the high activity tiles in a metropolitan area where there is dense WiFi but that are currently underserved by fiber?
Activity graphing for time of day and time of week to optimize placement of crowd sensitive or dependent commercial or public properties.
Example: Where are the busy places with bakery customers where I should open my next Panera?
TIDE5 (or finer) discretization of where APs are detected during conventional traffic patterns vs during a storm surge, floods, earthquakes or other events highly disruptive to communication networks.
Example: Where are the places that have the highest value internet infrastructure (most functioning WiFi access points and cell antennas) during landfall of a major hurricane?
Tide Pool/tiled device counts/default distribution format (csv; partition=ymdh)
|time_ymdh||2017-08-17 14||UTC; timestamp to hour (local/umts)||ts|
|hour_of_week||167||Local time part; 12am-12:59am = 0
|hour_of_day||23||Local time part; 12am-12:59am = 0||int|
|latitude||42.351189||signed; 6 digits of precision .>||in-tile float|
|longitude||-71.040664||signed; 6 digits of precision .>||in-tile float|
|rlx_count||1618||count of returned locations logged within tile for hour||int|
|devices||64||count of confirmed devices within tile for hour||int|
|cateogry_id, source||301||Variable: signal source (os, app_class, stream)||int|
The above description outlines the default distribution of the TIDE product. Because time and space are essentially two knobs that Skyhook can use to produce finer or coarser mesh outputs, the hours and tiles can be delivered as more or less granular samples based on the requirements of any location-critical use case.