A consequence of our constantly connected lives is the expectation of collecting data cheaply, on demand, and at all times. What is often little understood is the scale of the physical infrastructure that is needed to collect, process, and analyse that data. This can lead to over-estimating the ability of those who monitor schemes to be able to report data on demand, often at short notice, and usually for a purpose well outside the previously agreed scope. This is why, if you anticipate on demand reporting, you need to build it into your scheme monitoring plan from the outset so that you can get a good idea of its resource implications.

Let us take, as a basic example, the humble traffic count. A staple of traffic monitoring for decades. Look at the most basic example of the traffic count using a manual classified count (MCC). To do this sort of work the following is needed:

• Relevant permits and permissions for a classified counter (who is trained to count and classify vehicles correctly) for the time period when the counter will be out on site.

• Ensuring that the counter has the relevant equipment with them to undertake the count. At its most basic, its a high-vis jacket, pens, clipboards and paper. Some still carry actual physical counters. More often, counts are made using an app either on a tablet or mobile phone

• Inputting the count data into a data terminal or computer, usually using a spreadsheet. This needs to be independently verified.

• Analysis using either a spreadsheet at its most basic, but increasingly statistical packages such as SPSS or R. Visualisation is also increasingly important.

Automatic Traffic Counts (ATCs) are slightly more mature than this:

• Relevant permits and permissions for a counting loop and detection box to measure the axle weight of vehicles passing over them, and classifying the data accordingly. This includes any necessary road closures to install the loops, particularly those embedded into the road.

• Counting support staff having access to replace cartridges that register the count data. When doing a Speed and Volume Count (S+V), this cartridge replacement can be every day.

• Manually inputting, or more frequently now downloading the count data to a laptop in a spreadsheet. This needs to be independently verified.

• Analysis using either a spreadsheet at its most basic, but increasingly statistical packages such as SPSS or R. Visualisation is also increasingly important.

But traffic counting has come on leaps and bounds in the last 10 years, with the increasing use of video surveys and artificial intelligence to intelligently identify different road users. But even this has its difficulties:

• This equipment needs its own stand alone equipment, or be attached to a peice of street furniture such as a street light. This requires standard highway permissions for installation, as well as approvals from lighting technicians (especially where there is a power supply needed).

• Time needs to be taken to calibrate the camera or sensor to ensure that it is adequately detecting different road users, taking account of things like sunlight and dark skies.

• An adequate mobile phone signal is needed to transfer the count data to a data processing centre for analysis. Even then, some cameras maintain a cartridge back up in case of loss of signal.

• All data needs to be independently verified. This is increasingly done through AI, but may also need an independent verification from a human.

• Analysis is usually done through a bespoke platform constructed by the sensor equipment provider. But all provide for analysis using either a spreadsheet at its most basic, but increasingly statistical packages such as SPSS or R.

And this is just traffic counts, one of the simpler things that we do. A vast array of physical infrastructure is needed, along with the associated permits, approvals, trained staff, and maintenance regimes in order to get the insight that is needed. Not to mention that once the infrastructure goes in, then it is really, really hard to change without causing major disruption. That is why it is critical that this work is planned consistently and to a common standard from the outset.

Our job as professionals is to design how we collect data with the view to how we want it to be used throughout the life of the project. Something so simple that its frustrating that it needs stating, but something that is often not appreciated in a world where data is on demand, all the time, and visualisation of that data can be easily done with some simple tools. And the expectation as to what is feasible also needs to be managed.

A final, and often under-recognised thing that needs considering is how this data collection fits into a wider approach to collecting and managing data across an organisation. Very often, highway authorities have a (sub-part) Local Transport Plan, but seldom have a good quality data strategy to support their work and improve services. National Highways has an excellent example of a good data strategy, if you are looking for one.

Why is this important? Very simple really. Collecting data improves outcomes, and demonstrates impact. Doing it consistently and to consistent standards, and in a co-ordinated way shows that people can have confidence in your methods. Not to mention that uncoordinated data collection duplicates effort, results in poor quality data management approaches that are more sensitive to leaks, and increases the cost of monitoring.

But too often this is not done at all. And in a world where the expectations around data analysis and showing results is sky high, this is a dangerous thing to do. While having a robust data collection strategy as part of your Monitoring and Evaluation is crucial, even more so is having a data strategy in place that sets the standard.