TelecommunicationsBI, AI and ML for the business
A cell's evenings of interference
Radio cell counters arrive minute by minute: SNR, received power, users and dropped calls. A condition isolates the evenings when SNR collapses while power rises, Layers aligns them on their start and the heatmap shows the hour it happens.
Illustrative scenario: it describes a typical case, not a customer project.
A cell works all day and collapses every evening. The complaints arrive before the explanation.
Cell counters pile up in millions of rows, and network dashboards show daily averages that hide the problem: a cell that is fine all day and degrades for a couple of hours in the evening, when traffic is highest.
In Muvia the counters arrive every minute and the analysis works on the signals, not on averages. A condition says what counts as interference, Layers shows how each episode unfolds, and annotations record the cause: anomalous propagation, a neighbouring cell switched off for works, a thunderstorm.
- Who it's for
- Radio optimisation engineers, the network operations centre, network quality managers and whoever answers coverage complaints.
In Muvia, step by step
Real product screens, recorded on a project with sample data.
The video · Radio cell counters arrive minute by minute: SNR, received power, users and dropped calls. A condition isolates the evenings when SNR collapses while power rises, Layers aligns them on their start and the heatmap shows the hour it happens.
What you get
- The problem averages hide
- Per-minute analysis shows the evening hours that a daily average flattens.
- One shared definition
- The interference condition is saved with the analysis: everyone counts the same episodes.
- Causes stay on record
- Annotations explain every episode, so the next one starts from what is already known.
- Before the complaints
- The interference window shows in the counters before it becomes a customer complaint.
For the technical team
How it is built in Muvia
- 1
Bring the counters into a project
The network management system's per-minute exports arrive over SFTP and are appended incrementally. Every source becomes a table and a query.
- 2
Define interference
In Analysis a condition says what counts: SNR below 6 dB with more than 70 users, for at least 10 minutes. Each episode is coloured on the SNR, power, users and drops panels.
- 3
Overlay the evenings
The Layers view aligns each episode on its start: you see how dropped calls respond to the SNR collapse. The distribution compares drops with and without interference.
- 4
Find the hour and note the cause
The hour-by-day heatmap shows the window; annotations record anomalous propagation, works and weather. A network dashboard takes the figures to whoever decides.
- The data it needs
- Cell counters every minute: SNR, received power, throughput, active users, dropped calls, radio resources in use
- The site and cell inventory, with sector, azimuth and tilt, from an Excel file
- The log of planned works on the network
- The team's notes on weather and propagation, as annotations
- Parts of Muvia used
More cases in this line
All use cases- Antenna retilt, before and afterTwo drive-test campaigns on the same route, before and after the antennas were retilted. Muvia compares them lap by lap and stretch by stretch: SINR, received power and throughput tell you whether the work paid off, and on which stretches.
- A suspect lot, contained in minutesMaterial lots, orders, serialised parts, sequenced shipments and the customer's vehicle numbers linked in one path. When a supplier flags a suspect lot, the exact scope to contain comes out in minutes, instead of a blind recall.
- Sequenced deliveries that never stop the customer's lineCar makers' call-offs, JIS and JIT shipments and delays in one dashboard, with OTIF per customer and a map of when late call-offs arrive. Logistics sees the risk of a line stop before it turns into a penalty.
Bring us a question you can't answer today.
We start from a real question your business has and walk the path from source to dashboard with your systems, not a demo dataset.