Connecting Africa: Nomanini, Google Cloud Platform, and BigQuery
Understanding worldwide mobile networks with BigQuery
Nomanini is a startup powered by Google Cloud that enables micro-transactions in various countries around Africa. I flew to Johannesburg to meet their CTO and learn how they use App Engine, Kubernetes, BigQuery and other Google Cloud Platform tools.
Nomanini builds point-of-sale systems to help remote villagers get access to prepaid airtime, prepaid electricity, and basic banking services like being able to withdraw or deposit cash.
Back in 2011, they started by focusing on selling prepaid mobile air time. Distributing boxes of scratchcards out into rural villages isn’t easy — so they created a wireless alternative.
Now people in the villages don’t have to travel far to go and get essential services like air time or electricity, and even banking services.
In this video, Dale Humby — Nomanini’s CTO — shares more about the custom devices they created and why they chose Google Cloud to power their backbone. Originally on App Engine, they now have most of their infrastructure running on Kubernetes.
And for their analytics they use… BigQuery!
Each of their devices reports back its own stats — thus Nomanini is able to track their status and what kind of connectivity they have. This information allows them to debug the status of the different cell networks — and even inform the mobile networks of problems before they notice them.
OpenCellID: Open data for 36 million unique GSM Cell IDs
OpenCellID is the world’s largest open database of cell towers.
As of October 2017, the database contained almost 36 million unique GSM Cell IDs. More than 75,000 contributors have already registered with OpenCellID, contributing millions of new measurements every day to the OpenCellID database.
Thanks to open data, Nomanini can look beyond their devices and check the daily reports of thousands of contributors around the world.
Around the world insights
For example, let’s see what are the most popular mobile radio technologies around the world, according to the OpenCellID tables:
GROUP BY radio
ORDER BY c DESC
UMTS is the most popular radio tech according to OpenCellID, while CDMA is the least. But they’re not distributed evenly around the planet. A quick check with Data Studio can show us the actual patterns:
We also visualized how many cell towers each country in Africa has, relative to its population:
Nomanini is able to join this data with the reports from their own devices.
For example, to visualize the state of each of their devices around Mozambique (with Redash):
For the full story, watch our video!