Beacons in New York City
Welcome to Beacons in New York City
NYU CUSP IoT Team, 2021
Beacons
Beacons are tiny devices that use the Bluetooth Low Energy (BLE) signals to periodically transmit signals to mobile devices within a short distance or proximity of the beacons. Compared to devices based on Global Positioning System (GPS), beacons provide more accurate location information and can be used for indoor location technology. Various types of beacons exist, which can be classified based on their type of Beacon protocol, power solution and location technology. Major protocols for beacons are iBeacon, Eddystone and AtlBeacon.
iBeacon: a beacon protocol released by Apple in 2013. It is the first beacon protocol in the market. iBeacon works with iOS and Android.
Eddystone: a beacon protocol released by Google.
AtlBeacon: a beacon protocol released by Radius Network.
UUID: The beacon using the iBeacon protocol transmits a so-called Universally Unique Identifier (UUID). The UUID is a string of 24 numbers, which communicate with an installed Mobile App.
Our Work
Methodology
Research
Technology and Manufacturing Companies
We began with conducting industry research to find beacon manufacturers and their product UUID formats. Our methods included internet research, media research, cold calls and product demos. On the technology side, we conducted research on SDKs and app-libraries to understand the information flow ecosystem between retailer, beacon device, app and the user. Our primary technology research method was through the internet.
Data Collection
Feasibility Checks and Testing
While the research was in progress, we parallelly tested out several methods for ensuring successful data capture, often undergoing several iterations to find solutions to problems encountered.
Large scale data collection
We have covered areas such as Lower Manhattan, Tribeca, Lower East Manhattan, Chelsea, Hudson Yards, parts of Midtown, SoHo, NoHo, East Village, DUMBO, Downtown Brooklyn, Cobble Hill, Brooklyn Heights, parts of Bedford Stuyvesant and Williamsburg.
Data processing
The data from the server is read and saved as ‘.json’ files, while GPS records are read from the ‘.csv’ files using Jupyter Notebook. The beacon dataset contains all beacon related information as nested list of dictionaries within 2 columns: ‘beacons’ and ‘ibeacon’. To get useful beacon information, we extracted and expanded those columns, and then merged the beacon dataset with GPS location datasets using timestamps. Consequently, we achieved a consolidated dataset with 27 columns containing all information about beacon signals detected and the GPS coordinates.
Analysis and Visualizations
To communicate the findings, we used several tools to build effective visualizations such as Google Maps JavaScript API, D3.js, Plotly JS, Mapbox and Tableau Dashboard.
Ethics and Privacy
Considerations
The implications for many individuals regarding their personal data being collected without their consent is huge. Data subjects must be informed about the collection and use of their personal data when the data is obtained.
Data Insights and Results
In a span of 65 days, a total of 2,555 minutes and 102 kilometers of data collection across 129 census tracts in Manhattan and Brooklyn:
No. of beacon signals detected: 346,887
No. of beacon devices detected: 10,357
iBeacon
by Apple
7667 beacons follow iBeacon protocol
203 unique iBeacon UUIDs
Eddystone
by Google
2664 beacons follow Eddystone UID protocol
17 beacons follow Eddystone url protocol
11 unique Eddystone Namespace IDs (Eddystone version of UUID)
Altbeacon
by Radius Network
9 beacons follow Altbeacon protocol
2 unique Altbeacon UUIDs
Unique Beacons Detected
Where We Found Them
Google Maps JavaScript API was used to map the beacon locations, using the package “gmplot”. The GPS location data is plotted as grey dots, whereas the Beacon approximate locations are plotted using red pinpoints. In order to avoid unnecessary overlapping of markers, we also combined adjacent markers into one based on 4-digit GPS. Through the overlap of both datasets, one can witness the scale of detected beacons compared to the scale of locations we covered. Upon hovering on the markers, information such as ibeacon uuid, street address and GPS locations can also be gathered.
Geographical Distribution of Beacons by UUIDs
Using Census Tract of NYC
To visualize the distinct UUID level statistics of detected beacon devices, a combination of D3.js, Plotly JS and Mapbox was used. This visualization combines a bar plot and a census tract level map of NYC, and interactively displays counts of every UUID found, mapped back to its manufacturing company (if identified), and displays on the map the locations it was found at.
Hover and Click on the bars to explore!
.png)
Beacons in New York City Team
Get to Know Us
Contact Us
New York University
Center for Urban Science + Progress
370 Jay Street, 13th Floor
Brooklyn, NY 11201