This paper sets out the steps we have taken in first simulating a spatial dataset and then developing a model to estimate the capacity of people to travel actively in different areas of Scotland. This model takes into account relevant characteristics of the population including age, gender and health.
Spatial microsimulation is a well-established technique that estimates what individuals are like in an area, based on aggregate statistics. Our approach is based on similar recent work conducted on English data. In Scotland our source of health data was the 2016 Scottish Health Survey and our spatial data source was the 2011 census which provides data at output level (geographical areas of around 125 households).
We identified attributes that are required to model an individual’s capacity to travel actively and which are collected for the Scottish Health Survey, such as age, gender, whether the individual had a long-term limiting illness and whether he or she was economically active. The spatial microsimulation approach then allows us to combine these attributes with the spatial data to estimate the number of individuals in each category in each output area.
Our approach also incorporates data from the 2016 Scottish Household Survey to help simulate which individuals in our local populations have access to a bicycle – a critical factor in their ability to travel actively.
Using the simulated population, our model calculates a maximum cycling and walking distance for each individual based on a number of factors: an estimate of their VO2max (the maximum amount of oxygen that an individual can make use of during exercise) which, when combined with information on weight and BMI produces an estimate of an individual’s power output.
One of the other factors affecting the distance an individual can walk or cycle is the topography of the area in which they live. Our model takes into account the average slope of within five kilometres of the centre of each output area when calculating the cycling and walking speed for each individual.
We believe our model has produced a rich dataset with a variety of uses. For example, it provides a clear and easily communicable presentation of locations where individuals have relatively limited capacity to travel by active modes, both within and between areas. For example the mean distance that people have the capacity to cycle ranges from 10. 7 km in Inverclyde to 14.5 km in Angus.
Our modelling has also demonstrated that the common transport planning assumption that people can cycle 5 miles (8km) overestimates the capacity of the population to commute by active modes. We found that 21% of the individuals in our model did not have the capacity to cycle this far, while an even higher proportion had the capacity to cycle at least this far but did not own a bicycle. This highlights the risk of excluding substantial proportions of the population when relying on common assumptions.
There are a number of areas in which the methodology outlined in this paper could be improved and we intend to take these forward in the coming months. Nevertheless we are sharing our methodology at this stage to demonstrate the power of the approach and in the hope of inspiring further joint development in using it to model active travel behaviours in Scotland.
CORE (COnnecting REpositories)
CORE (COnnecting REpositories)