Each month we process millions of rows of information relating to our broadband network. This data is stored in multiple database systems and results in several hundred GB of data in its raw form. We then feed this data into several other systems for further processing. This data is used for troubleshooting, enforcing broadband caps and it also used by our business customers for monitoring their own usage.
The data we collect includes information such as the amount of time spent online and the volume of data downloaded/uploaded. We do not collect any data relating to the type of content or the content itself.
We thought it would be fun to do some basic data analysis using our anonymous usage data.
As a starting point, we have broken the Isle of Man broadband usage data down by post-code area. In time, we could introduce other variables such as the weather and public holidays. In the future, and following additional analysis, we should be in a position to answer a number of very important questions :
- Why do people in Maughold download more than Port St Mary?
- What region of the Island spends more time browsing?
- Does the Isle of Man weather impact browsing habits?
I have included an example showing “average” download amounts “by user” for Sunday 22nd December 2013.
In time, we may produce a “real-time” broadband heat-map. This would simply be in the form of a coloured map. This would show real-time activity on our broadband network by post-code area. This may show visual patterns relating to the weather, power-cuts, (rare) network outages and similar trends.