All work

BJ Bison

Telemetry from 600+ telecom tower sites, turned into fuel, availability and anomaly reporting.

When
Aug 2025 to Jul 2026
Kind
Telecom tower operations
Role
Operations data analyst
Stack
Galooli, OfficeTrack, Kriterion, Excel

Nigeria's mobile networks run on towers most people never notice. American Tower Corporation (ATC) builds and owns thousands of them, and operators such as MTN and Airtel rent space on them for their equipment. BJ Bison manages sites for ATC. For a year I was its operations data analyst for more than 600 of them across Imo, Ebonyi and Akwa Ibom, reading the telemetry and field data and turning it into the reports the operation ran on.

IoT-connected tower sites
600+
states: Imo, Ebonyi and Akwa Ibom
3
site classes: Silver, Gold, Platinum and Colo
4
of fuel anomaly findings, with a workbook the team adopted
20 pages
On this page
  1. How a tower site works
  2. My role, and the hard part
  3. Fuel that doesn't add up
  4. Proof that someone was there
  5. From device data to decisions
  6. In use: a flag, from telemetry to action

How a tower site works

Think of it as real estate. ATC is the landlord: it builds the site, with its tower, shelter, generator and fuel tank. The operators are the tenants: MTN, Airtel and others lease space on the tower and in the shelter, under agreements that set how the site must be kept. BJ Bison is the facility manager, running each site to ATC's directives, which follow from those agreements.

No two sites are the same. They differ in size, in priority and in class: Silver, Gold, Platinum, or Colo for a site shared by more than one operator. Each is run to the terms its tenants agreed.

  1. ATCBuilds and owns the site
  2. OperatorsMTN, Airtel and others lease space
  3. BJ BisonRuns the site to ATC's directives
  4. Field teamsVisits, refuelling, repairs
  5. The dataGalooli, OfficeTrack, Kriterion
  6. ReportsWeekly and monthly, to operations
Who does what at a tower site

My role, and the hard part

I read the telemetry of every connected site in my three states, reconciled it with what the field teams logged, and turned both into reporting the operation could act on.

The hardest part wasn't the data. It was getting instructions to the field engineers and making sure they carried them out exactly. A directive starts in ATC's agreement with an operator, and by the time it reaches an engineer at a site in Ebonyi it has passed through several hands. A step lost along the way means a missed check, a wrong reading, or a site run below the standard its class demands. So I wrote instructions that couldn't be misread: the site, the task, the exact steps, and what to log as proof it was done. Then I checked the field data against them, so a skipped step showed up in a report before it showed up as a fault.

Findings travelled the other way too. A number on a dashboard means nothing until the right person acts on it, whether that's a field engineer, a diesel vendor, an operations lead or ATC itself, so every one said what happened, at which site, what the data shows, and what should happen next.

Fuel that doesn't add up

I built a fuel-consumption and theft anomaly analysis that flagged gaps between the diesel vendors said they delivered and what sites actually burned. It became a 20-page report and a reusable Excel workbook the operations team adopted for ongoing monitoring.

Proof that someone was there

An OfficeTrack field compliance report reconciled logged site visits and refuelling against the expected schedule, giving operations a clear view of which vendors and field teams were keeping to it.

From device data to decisions

Reporting pipelines turned raw device data into weekly and monthly consumption, availability and anomaly reports for operations leadership. Before any of it reached a report it passed data-quality checks for completeness, timeliness and consistency across sources, and I worked with field engineers, vendors and operations leads to validate the findings and act on them.

In use: a flag, from telemetry to action

A generator's fuel level in Galooli rises by less than the litres a vendor logged as delivered. The weekly workbook flags the gap against the site's usual burn. OfficeTrack shows the refuelling visit logged, but at an hour when the telemetry saw no rise at all.

The finding goes to the operations lead with the site, the dates, the numbers from both systems and a recommendation, and the vendor is asked to account for the difference.

An illustration of how a finding moves. No real site or figure is shown: the data belongs to the client.

Blockchain for IoTNext project: Journal paper, first author