Connecting Africa: Nomanini, Google Cloud Platform, and BigQuery

Understanding worldwide mobile networks with BigQuery

Watch the full video here

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’s challenge

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.

In the video, we use a combination of BigQuery, Data Studio, and re:dash to visualize the status of cell networks around the world, and in the cities where Nomanini is present.

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:

SELECT
   radio, 
   COUNT(*) c
FROM `fh-bigquery.public_dump.opencellid_201803` 
GROUP BY radio
ORDER BY c DESC

The most popular mobile radio technologies around the world, according to the OpenCellID tables

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:

CDMA is primarily seen around the US, while the rest of the world uses GSM

Fourth-gen networks like LTE are seen in Africa, but mostly in the big cities. Asia shows huge adoption.

We also visualized how many cell towers each country in Africa has, relative to its population:

Country ranking by cell towers per capita

Cell towers per capita in each African country

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):

Next steps

For the full story, watch our video!

Video: Connecting Africa: Nomanini, GCP, and BigQuery. Press play to watch.

OpenCellID is an open data project — let us know if you’d like to see it permanently updated in BigQuery through our public datasets program.

Want more stories? Check my Medium, follow me on twitter, and subscribe to reddit.com/r/bigquery.

And try BigQuery — every month you get a full terabyte of analysis for free.

Can we process 10.6 billion rows with Dataflow while on a tram? Check us out.