SignalMap
Import a CSV with coordinates and signal measurements to start.
| # | Timestamp | Latitude | Longitude | Technology | Signal value | Quality | Location | Map |
|---|---|---|---|---|---|---|---|---|
| No measurements yet. Import a CSV to view your data. | ||||||||
Import measurements to see the quality breakdown.
Groups nearby weak signal measurements into approximate spatial clusters. Results follow your selected metric and technology.
Clusters show where weak measurements were recorded, not coverage predictions for places without measurements. Grid connectivity and priority labels are approximate.
Compare the active CSV against a reference CSV, using matched spatial cells and the same signal metric / technology.
No reference CSV loaded. Changes are grid-based averages, not paired drive-test readings. Improvement means stronger signal (or lower latency) on matched cells; sampling differences may affect results.
Interpret 4G radio quality as a transparent 0–100 index. Only measured RSRP, RSRQ and SINR values are used.
Indicative formula: RSRP (-120 to -80 dBm) 40%, RSRQ (-20 to -7 dB) 30%, SINR (-5 to 25 dB) 30%; capped to 0–100 and reweighted when a metric is missing. Not an official operator QoS score. Cells indicate measured sample groups, not predictions or percentage of geographic coverage.
Animate the recorded measurement sequence on the map. An actual route requires the CSV to be exported in acquisition order.
Sequence follows current metric and technology filters, preserving CSV row order. If rows were not collected sequentially, the connecting lines do not represent a real driven route. Timestamps are shown when available but are not inferred or synthesized.
Draw a polygon on the map to analyze recorded samples inside it. No interpolated or unsampled areas are counted.
Weak share represents the share of measured points in your polygon that meet the selected metric's weak threshold. It does not estimate the fraction of the territory without signal.
Indicators from your measured RSRP, RSRQ and SINR values. These are possible causes, not confirmed root causes.
Confidence means measurement density within sampled grid cells, not guaranteed coverage between measurement locations.
Save measurements, reference dataset, filters, map view, and drawn polygon into a portable JSON file. No cloud storage.
Your existing 5-page OCP SignalMap Report is preserved. Optionally include a technical evidence appendix using Diagnostics, Confidence Map, CSV Validation and selected area results.
Review parsed rows, missing coordinates, out-of-range measurements and exact duplicates before activating a dataset.
The original dataset stays on the map until you confirm a validated replacement. Processing happens in your browser.
Discover serving cells, PCI use, network bands and cell changes using uploaded data only. No inferred tower coordinates.
Import CSV containing Cell ID, PCI and/or Band. Identifiers that are absent will not be inferred.
Switches count changes in the original CSV row order, not verified handovers. The same PCI may legitimately be reused by different cells.
Split-map inspection of actual samples from a reference campaign versus the current campaign.
Import a reference CSV using the Compare tab to activate.
Left and right layers share the same georeferencing. They depict sampled locations only; they do not interpolate unmeasured coverage or imply spatially matched changes.
Measured RSRP, RSRQ and SINR by CSV order or accumulated GPS distance. Optional smoothing is explicitly labeled.
Import measurements to draw RF charts.
GPS distances are derived from consecutive recorded positions and may include jumps.
Organize follow-up actions for detected weak clusters. Notes and statuses are editable, not automatic diagnoses.
Tasks are stored in current page memory and optionally in a downloaded Project JSON. Device save is voluntary; do not use local storage on a shared computer for sensitive coordinates.
Switch to a local-only field grid, keeping coordinates, recorded points, heat overlays, analysis and Project JSON available.
3D terrain studio with interactive orbit, local RF color overlay, and existing 2D scientific model. Synthetic terrain is not actual elevation unless imported.
GPS of the base station — not measured route points
Waiting for site coordinates.
Editable assumptions; a propagation approximation, not a planning-grade ray tracer
Reference: ITU-R P.525 FSPL at 100 m; log-distance propagation beyond 100 m; simplified sector attenuation. The offset converts a received-power estimate to an RSRP proxy and is not standardized.
Baseline remains fixed while you adjust the candidate
Azimuth · downtilt · transmit power
Optional median bias correction · RSRP records only
Calibration is disabled until you confirm a verified antenna site. Use matching-cell RSRP measurements where possible. A 20% holdout is used to show predictive error.
Orbit around a 3D surface with the same empirical RF model as the 2D simulator
Real-world elevation from DEM tiles, 3D satellite/street map and georeferenced antenna. No fabricated mountains.
Every identifiable site and cell. Physical towers stay anchored to explicit site coordinates.
| Select | Site ID | Cell ID | PCI / PSC | Operator | Technology | Coordinates | Location confidence | Measurements | Mean RSRP | Source | Classification | Actions |
|---|
Blue IAM · purple inwi · orange Orange · gray unknown. Shared sites retain all operator evidence. Coordinate-known does not mean independently surveyed. PCI/PSC alone cannot identify a unique cell. Public OSM records keep a separate classification. Dashed observation groups are not tower locations or predicted coverage boundaries.
Matching preserves source coordinates and categories. A nearby public tower is never matched automatically.
Modeled received-power proxy, separate from measured coverage. Explicit radio parameters are required. Log-distance loss anchored to free-space loss at 100 m; fixed 65° horizontal / 10° vertical beamwidth assumptions. No interference, handover, buildings or terrain propagation.
Search real mapped tower/mast coordinates near the map — independent of your CSV. Green = tagged mobile, blue = telecom/radio, brown = unverified mast/tower. Not guaranteed to include every antenna.
No external data loaded.
© OpenStreetMap contributors · Overpass API. Public structures use separate markers; unknown heights remain unavailable. Requests send only geographic search bounds, not CSV measurements.
Choose a terrain source and generate a 3D scene.
Import CSV with Latitude, Longitude and Elevation_m. Surface is interpolated between samples (not a validated DEM).
No elevation file loaded0 vertices · 0 triangles · waiting
Cell color = modeled signal, not satellite imagery
No scenario yet. Run RF simulation.
Click a square for GPS, baseline, candidate and Δ. Map overlay updates without replacing measured points.
Transparent equations and uncertainty
Horizontal/vertical attenuation is a simplified sector pattern based on angular distance from azimuth and downtilt. Free-space attenuation is a reference, not a terrain-aware prediction.
Unmodeled: topography, urban clutter, interference, handover, traffic loading, bandwidth, exact reference-signal EPRE, multipath and atmospheric details.
Calibration: reported bias uses only training samples. Holdout MAE uses unseen imported rows; it does not establish performance outside measured locations.
Technical references: ITU-R P.525-5 (free-space attenuation); 3GPP TR 38.901 (antenna patterns, conceptual guidance).
Run self-checks using synthetic CSV values without overwriting measurements, reference campaigns or map settings.
Self-checks validate calculation contracts and essential loaded dependencies. Complete end-to-end CSV imports, real map rendering and PDF save must also be tested in a browser with those libraries available.