Small Pings, Big Decisions

How AVL ping data can become a sustainable source of transport insight

Tom Alexander

From Pings to Insights

What I want to cover:

  1. Why are we looking beyond existing tools?
  2. What insights can AVL ping data provide?
  3. How do we build a sustainable analytics capability?

Every Few Seconds…

Every few seconds, buses tell us where they are.

What can we learn from millions of tiny pings?

Where Are We Today?

ABOD - DfT

Icarus (from our RTI supplier)

Can We Get Somebody Else To Do It?

Option 1:

Ask the DfT

  • Enhance ABOD
  • Add new functionality
  • DfT have said “no”

Option 2:

Buy Something

  • Commercial Data Product
  • Limited flexibility

Option 3:

Build It Ourselves

  • Develop analytical pipelines
  • Answer the questions that matter to us
  • Adapt as our needs evolve
  • we’re not used to do doing the analysis ourselves
  • resource and commitment
  • thinking longer-term

The challenge isn’t the technology. The challenge is creating the time, space and capacity to do it properly.

Data Engineering Foundations

✅ Captured AVL data from the BODS API

✅ Stored millions of location records in compact, efficient parquet files

✅ Developed processes to clean and standardise the data

✅ Established a repeatable workflow for handling incoming data

Last Monday…

“Have you got any evidence to help prioritise schemes within the TCR programme?”

Tom Campbell,

TCR Programme Manager

a real business question was translated into useful evidence in just over a week

A Project Manager’s Question

We already know which corridor we want to improve. Where along the route will intervention have the greatest impact?

Future Idea 1: Explaining Delay

Combine

  • AVL Ping Data

  • On-bus Delay Categorization Survey

What can the survey add?

  • Help us to understand why journeys are slow.
  • proportion of a journey time that is delay by:
    • traffic or
    • passenger boarding etc

Power of combining other data sets

Combine:

  • TomTom traffic speed data
  • Patronage Data
  • demographic data

we can sart to unravel other questions:

  • is bus priority working?
  • would alternate vehicle design speed up passenger loading?
  • Does stop design affect operational performance?

Key Takeaways

✅ Existing tools don’t answer all of our questions

✅ AVL data can support real project, operational and strategic planning decisions

✅ Already providing supporting evidence for programme decisions

✅ The most valuable insights will come from combining datasets

✅ A ‘skeleton’ analytical pipeline already exist.

What do we need next?

Continued development of analytical pipelines :

  • Data engineering support

  • Analytical support

  • Business analysis to identify priority questions

Small Pings. Big Decisions.
A small commitment now unlocks a long-term source of transport insight.