Dataset used in "Balloons in the Sky: Unveiling the Characteristics and Trade-offs of the Google Loon Service"

This dataset was used for the paper "Balloons in the Sky: Unveiling the Characteristics and Trade-offs of the Google Loon Service"

It is composed of three different files, identifying three different case studies Peru, Puerto Rico, and Kenya

Once uncompressed, the files are stored in parquet format. They can be analyzed using  e.g. pyspark with the following code

df = spark.read.parquet('dataset.parquet')

Each scenario is composed of two tables.

The file dataset.parquet contains the measured bandwidth in the analyzed area for a period of approximately three months. Its columns are:

• ts: the timestamp of the measurement, as UNIX timestamp
• h3: the id of the H3 hexagon (aka pixel) of the measurement 
• nloons: the number of balloons that provided coverage to the pixel at the given timestamp
• thr: the available bandwith
• serviceDuration: the duration of the service on the given pixel, starting from ts

The file serving_loons.parquet contain the information about the available balloons flying over the area under study. Its columns are

• ts: the timestamp of the measurement, as UNIX timestamp
• serving_loons: the number of ballons flying over the area