Introduction

The following analysis looks at travel time to various locations in Glasgow and how these vary spatially and by socio-economic factors i.e., sex, age, ethnicity, and deprivation. It also looks at how accessibility to jobs in different industries varies.

By public transport we mean buses, the subway, and trains. Bus and subway data is taken from Traveline. It does not seem to cover services such as the airport bus. Rail data is taken from the Rail Delivery Group. Traveline and rail data refers to the first week of September 2020. Historic data does not seem to be available to allow us to see the situation pre-lockdown. It is probably available somewhere but not through the same channels. It will be relatively straightforward to rerun the analysis with updated timetable information. Journey times are likely less affected by disruptions due to the lockdown, however wait times may be longer due to reductions in the frequency of services.

I have assumed that journeys are taking place at noon on a Monday. Travel times will be vary depending on the time of the day and the day of week. We can perform calculations for whatever day and time is of interest. Calculations are based on timetables reported by the operators rather than actual performance.

Data from Traveline was supplied in TransXchange format while data from the RDG was supplied as ATOC CIF files. These formats do not work with standard tools. They were therefore converted to General Transit Feed Specification (GTFS) using UK2GTFS R package. The GTFS data was them combines with OS Open Road data to form a network dataset. This was done using the Add GTFS to a Network Dataset add-on tool for ArcMap 10.6. The Network Analyst extension was then used to conduct the network analysis. Ideally this workflow will eventually be moved to an open platform such as Open Trip Planner

Travel times are measured based on population-weighted data zone centroids. The population weighting is based on the 2011 census. Some parameters of the analysis can/should be tweaked e.g., walking speed, time taken to board/alights. The default parameter choices should, however, give a reasonable indication.

Some caution should be exercised. The timetable data can be messy and the UK2GTFS package is still under development. Furthermore, the Add GTFS to a Network Dataset tool has been deprecated as its functionality has been incorporated into ArcGIS Pro. However, the results look mostly reasonable. We can check them against other sources to ensure robustness.

Travel time in Glasgow

Below, we can see the travel time to different locations in Glasgow.

Population

Below we can see how many people live in each of Glasgow’s data zones. There is not such a large amount of variation because data zone boundaries are drawn to have roughly equal populations.

Travel time by sex

For the analysis of how travel time varies by sex, we begin by looking at how the split between male and female varies across the city. The map below shows the percentage of the population who is male according to the 2011 census. The main pattern is that there is a higher proportion of males living in and around the city centre.

Next we compare the travel time to different locations in Glasgow for males and females. This is done by taking a weighted-average of the travel times from data zone centroids across the study area. Women tend to have longer travel time to all destinations. This is driven by the fact that there are quite a lot of males living in the well-connected city centre compared to females.

# Travel time and deprivation

Next we move to look at deprivation, beginning by plotting deprivation quintiles. The first quntile is the most deprived. The quintiles are based on rankings within Scotland rather than only within Glasgow.

As before, travel times for each category are calculated by taking weighted averages. They are plotted below for each combination of destination and deprivatoin quintile. The travel time in minutes is also shown.

Travel time by age

The maps below shows the distribution of people by age group in Glasgow. The main pattern is that young people tend to be more likely to live in the city centre. The effect is particular noticable for people aged 19-25. Presumably this is the student population. The effect persists somewhat in to the 26-35 age group.

The first plot below shows the average travel time to destinations in Glasgow based on the age categories used on the above map.

Below we plot the data in more detail, using one-year age groups. Notice the dark, vertical band around the 19-25 age group. This indicates lower travel times to all destinations. Again, this seems to be driven by the tendency of this age group to live in the city centre.

Travel time by ethnicity

Below we plot the number of people of different ethnicities (based on the 2011 census) in each of Glasgow’s data zones. We exclude white people as it is harder to see any other patterns if they are included. Some clear patterns are visible e.g., the cluistering of the Pakistani community on the South Side.

Below we see the average travel time to different destinations for different ethnicities.

Access to jobs

In this part of the analysis we look at accessibility to different jobs. The number of jobs is taken from Nomisweb and is based on the Business Register and Employment Survey (BRES). Data from 2018 was used as it is the most recent. There are many different ways to measure accessibility to employment. One of the simplest measures is a so-called isochrone measure, where the number of jobs within a travel time threshold are counted. The strength of this measure is that it is easy to interpret. It can, however, lead to some cliff edges e.g., where a large number of jobs lie just over the threshold they are not counted. For the calculations here I have used a threshold of 30 minutes. This can easily be changed

Another point to note is that I have included access to jobs within the Clyde Valley Planning Area. We may want to go beyond this if we expand the travel time threshold. For example, if we have a threshold of one hour then people living in Glasgow city centre would be able to reach jobs in Edinburgh. This can be incorporated quite easily.

The map below looks at how many jobs can be accessed within 30 minutes.

The map below is similar, however now we separate out jobs by industry. There are several industrial classifications which we could use. I have used this one as it doesn’t have too many categories but still seems to show some interesting patterns. This can be easily changed. The map shows which industries dominate Glasgow and where they are accessible from. Some idnustries are quite concentrated while others are more spread out.

Below we look at how many jobs can be accessed in 30 minutes by one-year age group. I have restricted the plot only too show people aged 16 and over. We see that people in their early 20s tend to have the best accessibility to jobs. Again, this is due to them living in the city centre.

Below we look at how job accessibility varies by ethnicity.

Below we see how access to jobs varies by deprivation. There is a clear tendency that people living in the most deprived neighbourhoods have the lowest access to jobs. The accessibility improves as we look at less deprived neighbourhoods.

We can see how the accessibility to jobs varies by sex in the chart below. Males tend to have better accessibility to jobs. Once again, this is most likely driven by the concentration of males living in the city centre.