Near / Far · Inverse Distance Weighting (IDW)
When a school, park, playground or cycling corridor has no dedicated sensor, CitySense estimates local conditions from nearby active sensors. Nearer sensors receive more weight; farther sensors receive less.
Formula used
Estimated value = Σ(wᵢ × xᵢ) / Σwᵢ
wᵢ = 1 / max(dᵢ, 50 m)
This is IDW with exponent p = 1. The 50 m floor prevents a very close sensor from receiving an excessive weight.
Progressive near → far search
The search starts within 500 m, then expands to 1000 m when needed. For parks and cycling corridors, the limit can reach 1500 m.
At most the 4 nearest active sensors are used. If none is available inside the methodological radius, no estimate is produced.
How to read the platform correctly
Cluj CitySense is not meant to replace official monitoring. Its main role is to reveal local differences, trends and urban context that are hard to see otherwise.
Where the data come from
Hardware in use
The CitySense network will use nodes with Sensirion SEN50 and SEN55, both from the SEN5x family.
SEN50 provides PM1, PM2.5, PM4 and PM10. SEN55 adds temperature, relative humidity, VOC Index and NOx Index. Noise and UV remain available on configurations that include dedicated components.
What that means in practice
A single device measures several parameters in the same place.
This helps describe the local microclimate and the everyday urban experience around that location.
How the map is built
The basemap is only a visual support. The coloured layer on top is calculated from readings of active sensors.
The heatmap is built locally around each active sensor. Values are no longer interpolated across large areas between distant sensors.
To avoid unrealistic results, each sensor is limited to an influence of about 420 m around the measurement point; one sensor should not colour the whole city.
How freshness and quality are handled
Freshness filter
Sensors without a valid reading in the last hour are removed from the map, rankings and calculations.
This reduces the risk that stale data create a false sense of active monitoring.
Approximate values for schools, parks and playgrounds
If a place does not have a sensor directly installed there, the platform may show a geographic average of the nearest active sensors.
That estimate is clearly labelled as an approximation.
Near / far estimation: Inverse Distance Weighting (IDW)
When a school, park, playground or bike corridor has no sensor installed directly at that location, CitySense estimates local conditions from nearby active sensors. The method is a form of Inverse Distance Weighting (IDW).
How it works
Closer sensors receive more influence than sensors farther away. For each parameter we calculate a weighted average:
estimated value = Σ(wᵢ × xᵢ) / Σwᵢ, where wᵢ = 1 / max(dᵢ, 50 m).
The 50 m floor prevents an extremely close sensor from receiving an unrealistically dominant weight.
Progressive search radius
The platform first searches within 500 m. If there are not enough local data, it expands to 1000 m; parks and bike corridors may expand up to 1500 m.
At most 4 nearest active sensors are used. If no sensor is available within the methodological radius, the location is shown as having no local coverage rather than inventing a value.
What the indicators mean
PM2.5
Shows fine-particle concentration in air, expressed in µg/m³.
It is useful for spotting the effects of traffic, combustion and weather conditions.
Noise
Displayed in dB(A), it describes sound levels in a way that roughly follows human hearing.
It is strongly influenced by traffic, worksites, crowds or quiet streets.
Temperature, humidity and UV
These values describe local microclimate and urban comfort.
Shade, vegetation, sun exposure and mineral surfaces can change them a lot.
VOC Index and NOx Index
On SEN55 nodes, CitySense can display relative indices for volatile organic compounds and oxidizing gases/NOx.
VOC Index is relative to recent history, with the standard average condition mapped to 100. NOx Index has a standard baseline of 1; increases indicate events. These indices are not direct concentrations in ppb or µg/m³.
Urban score
The 0–100 score is a CitySense indicator.
It combines several parameters into an easy-to-read summary of the overall state of a place.
Accuracy and limitations in simple terms
PM particles · SEN50 / SEN55
For PM1 and PM2.5 in the 0–100 µg/m³ range, the SEN5x family used by SEN50 and SEN55 specifies ±5 µg/m³ and ±5% of the measured value.
That is why it is not correct to reduce the full specification to just “±5%”.
Temperature and humidity · SHT31
For SHT31-DIS-B, the typical specification is about ±0.2 °C and ±2% RH under the manufacturer’s conditions.
In the field, housing, airflow and solar radiation can modify the response.
VOC / NOx · SEN55
With Sensirion’s standard algorithms, VOC Index and NOx Index are relative indices, typically on a 1–500 scale. VOC uses a standard average condition mapped to 100, while NOx uses a standard baseline of 1.
CitySense does not present them as direct TVOC or NO₂ concentrations without a separate conversion and validation method.
Recommended technical sources
CitySense ecosystem
The Cluj implementation is synchronized with the umbrella CitySense methodology, and bicycle geometry is served centrally so mobile and desktop use the same network.
Colour palettes used across the platform
Help pages, legend, markers, rankings and bike lanes use the same colour functions.