Device Location
Device Location refers to the device_location table that contains our raw location data after it’s been de-identified, processed, cleaned, and normalized.
The device_location table is the foundation of all our core data assets and is used to build Stops, Visits, Trajectories, Device Metrics, and Density Tables. The most relevant fields in the device_location table are:
-
- Device ID
- Latitude
- Longitude
- Timestamp
Stops
Stops represent the points at which devices spend time, and are built according to the spatial and temporal proximity of individual device pings. Our default threshold for a Stop is a Dwell Time of 2 minutes, but clients can change this to suit their needs.
Dwell Time
-
- Dwell Time represents the amount of time a device spends at a location. We collect on average over 100 data points/user/day and use this data to determine how long users spend at different places.
- Dwell Time also helps us distinguish actual visits from non-relevant data points. For example, just because a device pings near a movie theater doesn’t mean that person saw a movie – they could have just been walking by, which is why you need to consider how long the device spent at that location.
POI
POI refers to ‘Point of Interest. They’re often physical stores, but can be a hospital, university, corporate building, park, or residential area. POIs are characterized by attributes such as polygon coordinates, characteristic dwell time, opening date, and opening/closing hours.
Sometimes POI refers to ‘Polygon of Interest’ – or, an Area of Interest (AOI) that is not characterized by any attribute but its polygon coordinates.
Visit
Stop + POI = Visit
Visitation refers to the number of people who visit a store or business during a particular period of time. Spectus measures visitation to:
-
- Determine campaign effectiveness in driving consumers to a store
- Understand store/brand performance
- Map competitive share of visits and understand market share evolution.
Trajectories
Trajectories refer to the paths traveled by a user between two consecutive stops in a given day.
Trajectories are a specific, queryable table that businesses use for origin-destination analysis. Trajectories allow businesses to map the route, speed, and frequency of travel between points over a specified period.
Density Metrics
Density Metrics refer to the different metrics businesses can use to analyze the density of devices in an area. Popular Density Metrics include the average, minimum, and maximum number of distinct devices in an area over a specified period.
In technical terms, Density Metrics refer to all possible metrics you can build starting from stoppers’ counts per geotile information. The stoppers_metrics% tables include these Density Metrics:
-
- Avg_daily_distinct_devices
- Min_daily_distinct_devices
- Max_daily_distinct_devices
Density Tables
Density Tables are pre-processed, queryable datasets for counting the number of distinct devices in an area over a period of time. Density Tables are useful for normalization, and can be queried on various geotiles over monthly or daily time periods.
stoppers_metrics% have a monthly granularity
stoppers_hll% have a daily granularity
Our most popular geotiles are bing tiles, geohash, and h3 cells. Learn more.
For more details refer to the Data Catalog app and Tutorials app
6 Density Tables are available in the paas_cda_v2 schema containing stoppers’ counts per geotile.
- stoppers_metrics_by_bing_tiles
- stoppers_metrics_by_geohash
- stoppers_metrics_by_h3
- stoppers_hll_by_bing_tiles
- stoppers_hll_by_geohash
- stoppers_hll_by_h3
II. Analyses
Longitudinal Footfall Analysis
Longitudinal Footfall Analysis produces the count of distinct devices that stop in, or pass by one or several polygons over a specified period.
Longitudinal Footfall Analysis helps answer questions like:
-
- How many people visited a POI in the last week?
- How many people stopped in 10 specific buildings over the last 2 years?
- Over the last month, how many people stopped in a polygon for at least 30 minutes?
Spectus provides 2 solutions for Longitudinal Footfall Analysis:
-
- API: For time sensitive use cases
- Jupyter notebook: Non-time-sensitive use cases, or when looking for footfall data for many predefined POIs
III. Technology
Differential Privacy
Differential Privacy is a mathematical definition of privacy and is achieved via a variety of mechanisms that add noise to the process generating the dataset itself.
Privacy Enhanced Mobility (PEM)
PEM, our patent-pending privacy enhancing technology, improves user privacy preservation, with 88% improvement over industry standards.
Privacy by Design (PbD)
Privacy by Design is a rigorous and internationally recognized framework designed to protect sensitive data by default. Spectus’ privacy strategy is grounded in the 7 principles of PbD.
