Newcastle Urban Observatory sensors

[joeytalbot]

There are a wide range of data available:

• counts for several vehicle types
• plates matching
• traffic flow
• average speed
• congestion etc

[joeytalbot]

Here are the sensor locations for car counts in 2021. They span across Tyne and Wear and beyond.

[joeytalbot]

Number plate matching:

[Robinlovelace]

Looking like a good basis for evaluating estimates of traffic volumes at the segment level :+1:

[joeytalbot]

2021 Newcastle data:

Car count:

• 207 point locations across Tyne and Wear
• 4000-8000 hours of data

Plates match:

• 311 matches (each a linestring containing 2 point locations)
• 100-360 days of data each

Traffic flow:

• only a small area around the Metro Centre

People count:

• covers 205 points across Tyne and Wear
• around 1500 hours of data

Walking:

• only a few points close to a city centre square

[joeytalbot]

Can get data raw, hourly, daily, weekly

[joeytalbot]

Plates In, Jan 2021

[joeytalbot]

[joeytalbot]

[joeytalbot]

After correcting the location data and removing bus sensors, the correlation is much better. R squared = 0.237. Still low, but the graph looks more reasonable.

[Robinlovelace]

Great stuff Joey!

[mem48]

Is the poor correlation due to commute data only? Would https://github.com/ITSLeeds/NTEM2OD/releases help?

[mem48]

Also, I've suggested to my new dissertation student working on traffic prediction to look into Newcastle as a possible case study location.

[Robinlovelace]

:+1:

[joeytalbot]

Days with full data; removed sensors with no traffic. R squared = 0.26

[joeytalbot]

[joeytalbot]

[joeytalbot]

[joeytalbot]

[joeytalbot]