## Comparing the densities of Australian and European cities

Thu 26 November, 2015

Just how much denser are European cities compared to Australian cities? And does Australian style suburbia exist in European cities?

This post calculates the population-weighted density of 43 Australian and European cities with a population over 1 million. It also shows a breakdown of the densities at which these cities’ residents live, and includes a set of density maps with identical scale and density shading.

As discussed in previous posts, population-weighted density attempts to measure the density at which the average city resident lives. Rather than divide the total population of a city by the entire city area (which usually includes large amounts of sparsely populated land), population weighted density is a weighted average of population density of all the parcels that make up the city. As I’ve shown previously, the size of the parcels used makes a big difference in the calculation of population-weighted density, which makes comparing cities difficult internationally.

To overcome the issue of different parcel sizes, I’ve used kilometre grid population data that is now available for both Europe and Australia. Some measures of density exclude all non-residential land, but the square kilometre grid approach means that partially populated grid parcels are counted, and many of these parcels will include non-residential land, and possibly even large amounts of water. It’s not perfect, particularly for cities with small footprints. For example, here is a density map around Sydney harbour (where light green is lower density, dark green is medium density and red is higher density):

You can see that many of the grid cells that include significant amounts of water show a lower density, when it fact the population of those cells are contained within the non-water parts of the grid cell. The more watery cells, the lower the calculated density. This is could count against a city like Sydney with a large harbour.

The second challenge with these calculations is a definition of the city limits. For Australia I’ve used Urban Centre boundaries, which attempt to include contiguous urbanised areas (read the full definition). For Europe I’ve used 2011 Morphological Urban Areas, which have fairly similar rules for boundaries. Both methodologies tend to exclude satellite towns of cities. While these boundaries are not drawn in the exactly the same way, one good thing about population-weighted density is that parcels of land that have very little population don’t have much impact on the overall result (because their low population has little weighting).

For each city, I’ve included every grid cell where the centroid of that cell is within the defined boundaries of the city. Yes that’s slightly arbitrary but I’ve been consistent. It also means some of the cells around the boundary are excluded from the calculation, which to some extent offsets the coastline issues. It also means the values for Australian cities are slightly different to a previous post.

All source data is dated 2011, except for France which is 2010.

### Comparing population-weighted density of Australian and European cities

You can see the five Australian cities are all at the bottom, most UK cities are in the bottom third, and the four large Spanish cities are within the top seven.

Sydney is not far below Glasgow and Helsinki. Adelaide, Perth and Brisbane are nothing like the European cities when it comes to (average) population-weighted density.

But these figures are only averages, which makes we wonder…

### How much diversity is there in urban density?

The following chart shows the proportion of each city’s population that lives at various urban density ranges:

Because of the massive variations in density, I had to break the scale interval sizes at 100 persons per hectare, and even then, the low density Australian cities are almost entirely composed of the bottom two intervals. You can see a lot of density diversity across European cities, and very little in Australian cities, except perhaps for Sydney.

You can also see that only 10% of Barcelona has an urban density similar to Perth or Adelaide. Which makes me wonder…

### Do many people in European cities live at typical Australian suburban densities?

Do many Europeans living in cities live in detached dwellings with backyards, as is so common in Australian cities?

To try to answer this question, I’ve calculated the percentage of the population of each city that lives at between 10 and 30 people per hectare, which is a generous interpretation of typical Australian “suburbia”.

It’s a minority of the population in all European cities (and even for Sydney). But it does exist. Here are examples of Australian-style suburbia in outer Hamburg, Berlin, LondonMilan, and even Barcelona (though I hate to think what some of the property prices might be!)

### How different is population-weighted density from regular density?

Now that I’ve got a large sample of cities, I can compare regular density with population weighted densities (PWD):

The correlation is relatively high, but there are plenty of outliers, and rankings are very different. Rome has a regular density of 18, but a PWD of 89, while London has a regular density of 41 and PWD of 80. Dublin’s regular density of 31 is relatively close to its PWD of 47.

### So what does the density of these cities look like on a map?

The following maps are all at the same scale both geographically and for density shading. The blue outlines are urban area boundaries, and the black lines represent rail lines (passenger or otherwise, and including some tramways). The density values are in persons per square kilometre (1000 persons per square kilometre = 10 persons per hectare). (Apologies for not having coastlines and for some of the blue labels being difficult to read).

Here’s Barcelona (and several neighbouring towns), Europe’s densest large city, hemmed in by hills and a coastline:

At the other extreme, here is Perth, a sea of low density and the only city that doesn’t fit on one tile at the same scale as the other cities (Mandurah is cut off in the south):

Here is Paris, where you can see the small high density inner core matches the high density Metro railway area:

Similarly the dense inner core of London correlates with the inner area covered by a mesh of radial and orbital railways, with relatively lower density outer London more dominated by radial railways:

There are many more interesting patterns in other cities.

### What does this mean for transport?

Few people would disagree that higher population densities increase the viability of high frequency public transport services, and enable higher non-car mode shares – all other things being equal. But many (notably including the late Paul Mees) would argue that “density is not destiny” – and that careful design of public and active transport systems is critical to transport outcomes.

Zurich is a city often lauded for the high quality of it’s public transport system, and it’s population weighted density is 51 persons/ha (calculated on the kilometre grid data for a population of 768,000 people) – which is quite low relative to larger European cities.

In my next post I’ll look at the relationship between population-weighted density and transport mode shares in European cities.

### All the density maps

Finally, here is a gallery of grid density maps of all the cities for your perusing pleasure (plus Zurich, plus many smaller neighbouring cities that fit onto the maps).

## Updates to transport trends – June 2015

Wed 10 June, 2015

I’ve recently updated three posts on this blog to include the latest available data. Here is a short summary.

### Transport greenhouse gas emissions

Australian domestic transport emissions have continued to rise and is now the sector with the biggest percentage growth since 1990. Domestic aviation emissions have tripled since 1990. Car emissions per kilometre were improving until 2007, but we appear to have gone backwards since then.

Full post here.

### Melbourne urban sprawl and consolidation

Outer growth areas of Melbourne now account for around 43% of population growth, but urban consolidation in the inner suburbs continues to exceed projections.

Full post here.

### Are Australian cities becoming denser?

This fully revised post looks at calculating population-weighted density using a new population grid for Australia, which finally allows for an internationally comparable measure of city density. I’ve also taken a look at some smaller Australian cities. The data suggests Sydney, Melbourne and Perth have been densifying fastest in more recent years.

Full post here.

## Trends in driver’s license ownership in Australia

Mon 9 March, 2015

Recent research has talked about “millennials” being less likely to get their driver’s license at younger ages, with data showing a decline over the 2000s. But is this trend continuing? This post checks out the latest data to see if the decline is still happening.

While I’m at it, I’ll look at license ownership by age and gender (are young men more likely to have a license than women?) and trends for older persons (are people holding onto licenses longer into old age?). There’s also a strange quirk for people born in 1945/6.

This post analyses available state-based data on driver’s license ownership in Australia in recent years. In this post some of the data sets I’ve used include learner’s permits, and some only count “independent” licenses (watch for notes).

### How does license ownership vary with age?

I have access to licensing data for four Australian states that allows a quite detailed analysis (three publicly, and VicRoads kindly let me access theirs).

The following chart shows license ownership for Victoria, South Australia and New South Wales by individual age for the most recent year available at the time of writing (combining licensing data with ABS state population estimates by single year of age to calculate ownership rates):

License ownership peaks between ages in the mid-thirties to late sixties, then falls away with age thereafter. There is certainly a pattern of people in their 20s and early 30s being less likely to have their license.

The main difference for the younger ages is that the Victoria data include independent licenses only (ie excludes people with a learner’s permit). Also, from what I understand, the minimum age for an independent driver’s license is 17 in most states, except in Victoria where it is 18, and the Northern Territory where it appears to be theoretically possible at age 16 and 6 months. The minimum age for a learner’s permit is 16 years, except in the ACT where it is 15 years and 9 months.

The Victoria data is generally higher than the other states from around age 32, with some results calculated as high as 99.8%. The Victoria data includes suspended licenses, which may not be the case for other states, and there may be other minor differences in the way the data is counted. But it is interesting that Victorian ownership rates are up to 10% higher for older age groups. I’ll look at those trends and patterns in more detail shortly.

I’ve previously looked at driver’s license ownership using VISTA data (Victorian household travel survey 2007-2009) which shows a similar pattern but with less detail:

### How does license ownership vary with gender (and age)?

First up, New South Wales:

Age 30 is the age under which females are more likely to have a license (or learner’s permit), and after which males are have higher rates of licensing. The difference in licensing between the genders grows very large for older ages. This might be explained by women of older generations being less likely to have ever obtained their license, and/or men stubbornly holding onto their license for longer than they should.

Here’s the same data for South Australia:

The gender flip occurs around age 27 – with younger women more likely to have their driver’s license.

Queensland data is available with slightly less age resolution:

The gender flip point occurs sometime between ages 21 and 24.

I really wasn’t expecting younger females to be more likely to have a license than males. Is this just something to do with learner’s permits?

I can only answer that question with Queensland data:

This shows women in Queensland are more likely to have their learner’s permit than men (at any age). However, men are actually more likely than women to have an independent license from age 20 onwards, as women appear to spend more time with their learner’s permit. It would be interesting to look at this for other states, but alas the data isn’t readily available.

You may also have noticed the South Australia and Queensland data suggests around 101% of men in some age groups have their driver’s license. This suggests imperfect data – perhaps double counting people with endorsements for higher classes of vehicle or people who have both car and motorbike licenses, or imperfect ABS estimates of people at each individual age. So licensing data needs to be read with caution, with a focus on the trends and patterns rather than exact numbers.

### License ownership trends of younger people

Here is independent license ownership rates in Victoria for younger people:

There are clear and sizeable downward trends in license ownership rates amongst most ages, with most dropping by around 12% over 13 the years. There was a slight rise in most age groups in 2009 but then a quite significant fall between 2009 and 2010, particularly 18 and 19 year olds. The graduated licensing system was introduced between January 2007 and July 2008, and I’m yet to find references to changes in rules around 2009 or 2010. so I’m not sure how to explain the changes in 2010. That said, when I look at the data for 2010, there are a few anomalies in patterns in other age groups, so there may be some small data errors.

The rate of license ownership of 18 year olds was however relatively steady between 2001 and 2008, but then dropped significantly from 2010. The minimum time period to hold a learner’s permit became 12 months in July 2007, making it harder to obtain your probationary license by age 18. There is a peak of license ownership at age 18 in (June) 2009 – these people will have turned 16 in the financial year 2006-07 and so probably escaped the new licensing regime (I suspect this cohort made more effort to get their learner’s permit before 1 July 2007).

I also note that the declines appear to have largely levelled off for most ages since around 2011. I don’t have much data about learner’s permit holders in Victoria, but some data published shows that the average time spent on L plates in Victoria for people aged 17-20 increased from around 60-70 weeks in 2000 to around 100 weeks in 2010, following the graduated licensing scheme introduction.

Have we now stabilised at new lower levels? More on that shortly.

Readers of this blog will note that several transport trends changed direction in 2011. That was about the time that public transport patronage growth in Melbourne slowed down, and mode share stabilised. Here’s the Melbourne mass transit mode share and young persons licensing rates charted together:

(mass transit mode share from other BITRE yearbook data)

Or if you look at this as a correlation:

That’s a strong correlation. Given that younger people dominate public transport patronage, this isn’t hugely surprising. The major deviation from the trend is 2009, which is perhaps explainable through changes to the licensing regime, although between 2001 and 2005 there was a reduction in license ownership without mode shift to mass transit.

So why has the trend towards lower license ownership of younger people stopped in Victoria? Other researchers might have to answer that question.

Here is data for young people in New South Wales (note: data includes learner’s permits):

Very different trends! Most age groups trended down between 2007 and 2010 but then many bounced up again thereafter.

So is this completely different trend because it includes learner’s permits? Unfortunately I don’t have single-age based data for people with independent licenses, but I do have it for age groups:

The trends are similar. Independent license ownership rates have dropped by around 3-5% in the younger age brackets over the nine years. There was a larger dip around 2009, followed by small rises in some age brackets since. Otherwise things look pretty stable, and very different to Victoria (and note that Sydney has had much less public transport patronage growth than Melbourne over the same time).

If I put learner’s and independent licences together and look at 20-24 year olds, there is a slightly higher proportion with their learner’s permit in more recent years (8% to 11%). So this suggests people are probably staying on their Ls for slightly longer.

Do the trends in NSW license ownership correlate with Sydney mass transit mode share?

Not nearly as much as for Victoria and Melbourne.

What is interesting in the NSW data is that there appears to be a pattern that varies by birth year. I’ve adjusted the layout of the data tables such that each row represents people in a single birth year (well, birth financial year, if you will). This data includes learner’s permits.

License ownership rates were relatively higher for people born 1992 onwards (although curiously they appeared to have declined after age 21, which perhaps might be a result of immigration – not sure). In NSW, the birth years of approximately 1982 to 1991 appear to have had relatively lower rates of license ownership.

Only six years of data is available for South Australia, but there is a pattern of higher license ownership for people born from around 1993 onwards (data includes learner’s permits), but it might be trending downwards again from birth years 1996 onwards:

(note: the only available South Australia data for 2010 is for January – I have interpolated to estimate June 2010 numbers. The most recent data is for March 2014 – I have interpolated the population estimate accordingly but in the figure above the bottom numbers in each column are for people born in the 12 months to March 1998, not 12 months to June 1998).

Queensland data only reports single year license ownership to age 20, but a much longer time series is available:

Again, there is a range of birth years from around 1984 to 1990 with relatively lower license ownership. In July 2007 the minimum age for a learner’s permit dropped to 16, which would explain the massive increase in learner’s permit ownership for 16 year olds. This seems to correspond to increased licensing rates for birth years 1993 onwards.

Although Victoria hasn’t had a bounce in license ownership rates for young people, the birth year trends do show the downward trend finishing with births around 1990. In fact, from that birth year the rate of license ownership at age 21 went up slightly, which might reflect that it is easier to obtain a probationary license from that age (as a learner logbook is no longer required). The decline in licensing rates seemed to begin around birth year 1980.

The following table summaries the birth years of lower license ownership in each state, and compares these birth years with the first birth year impacted by graduated by graduated licensing (assuming people obtain their learner’s permit at age 16):

 Lower license ownership birth years Graduated licensing started Start End Start First birth year New South Wales 1982 1991 2000 1984 Victoria 1980 1990 2007 1991 South Australia ? 1993 2005 1989 Queensland 1985 1990 2007 1991

There appears to be a fairly consistent cohort of people born between around 1980-5 and 1990-3 who have been less likely to get their license at a younger age. In Victoria and Queensland they weren’t faced with the new graduated licensing system if they got their learner’s permit at age 16. In fact, licensing rates stopping declining in the birth years first fully subjected to graduated licensing in Queensland and Victoria – the opposite of what you might expect!

Dr Alexa Delbosc at Monash University has led much interesting research into reasons for the downwards trend of license ownership in Australia (amongst others). Perhaps this new evidence of a reversal/stabilisation of the trend might explain some things further, or need a new explanation in itself.

First up, Victoria:

License ownership rates increased in Victoria until 2011 in most age groups, probably reflecting people living healthier for longer.  However license ownership rates fell from 2011 onwards for those over 80 (despite Victoria not having mandatory testing for older drivers). It might be explained by a change to a 3 year license renewal period for those over 75, but I cannot confirm when that change was implemented. I’m also at a loss to explain the blips in the 2009 data for those 90+.

Here’s the same for New South Wales, where car drivers need to have an annual medical review from age 75, and have to pass a practical test to keep an unrestricted license from age 85. For those 75+, licensing rates are considerably lower than in Victoria.

While the Queensland data provides less resolution of age groups, it shows the trend of increasing license ownership over a longer period of time, although with a levelling out for those 60-74 from around 2013.

And for completeness, here is five year’s worth of data available for South Australia, not showing any dramatic trends:

Finally, you might have spotted a blip in the first chart of this post around age 68 (you were looking at it carefully, right?). Zooming in and changing the X axis to birth year we see an interesting anomaly:

Years on the X axis are actually financial years (ending June). People born in 1946/7 are 10% more likely to have their driver’s license than people born in 1945/6, and this is consistent across three states!

This year of relatively lower license ownership is for people born immediately after World War II, and who around 19 years of age when Australia sent troops to the Vietnam War (although many men in the next birth year would also have gone to Vietnam). I wondered if it might be related to the Vietnam War, but then the trend applies more so to women than men, as shown in this NSW data:

I then thought it might be to do with being born whilst Australia was recovering from the war and healthy food and good medical care might have been less available, resulting in a mini-generation of people less likely to be able to get their driver’s licenses later in life.

The census provides one indicator of disability in terms of people recorder as “need assistance with core activities”. While people born between 1945 and 1949 seem to be slightly more likely to have a disability (compared to the general pattern across ages, supporting the post-war lower health hypothesis), the 1946 age year doesn’t stand out as much different to neighbouring years.

Another explanation might be that ABS have inaccurately estimated the population born in that year.

Can anyone else shed more light on this anomaly?

## What does the census tell us about cycling to work?

Mon 27 January, 2014

Who is cycling to work? Where do they live? Where do they work? How old are they? What work do they do? Do men commute by bicycle more than women? How far are cyclists commuting? What other modes are cyclists using?

The census provides some answer to these questions for the entire Australian working population, albeit for one winter’s day every five years.

This post builds on material I presented at the Bike Futures 2013 conference, using census data from across Australian with a little more detail on capital cities and my home city Melbourne.

It’s not a short post, so settle in for 13 charts and 17 maps of data analysis.

### How has cycling mode share changed over time?

The first chart shows the proportion of journeys to work by bicycle (only) in Australia’s capital cities.

Darwin appears to the capital of cycling to work, although it is quickly losing ground to Canberra (unfortunately I don’t have figures for Darwin pre-1996).  The census is conducted in Darwin’s dry season, but other data suggests there is little difference in bicycle activity between the wet and dry seasons.

Melbourne has shown very strong growth since 2001 and Sydney showed strong growth between 2006 and 2011. Cycling mode share has grown in all cities since 1996.

Mode shares collapsed in Adelaide, Sydney, Brisbane, and Melbourne between 1991 and 1996, which many people have attributed to the introduction of mandatory helmet laws (Alan Davies has a good discussion about this issue on his blog).

But as I pointed out at the start, census data is only good for one winter’s day every five years. Does the weather on these days impact the results?

Here is a chart roughly summarising the weather in (most of) the capital cities for 2001, 2006 and 2011 in terms of minimum temperature, maximum temperature and rainfall. It doesn’t cover wind, nor what time of day it rained (although perhaps some fair-weather cyclists might avoid riding on any forecast rain). It also fails to show the sub-zero minimums in Canberra (involves asking too much from Excel).

You can see that 2011 was wetter in Adelaide and Hobart than previous years, and this coincides with lower cycling mode shares in these cities in 2011. So census data is quite problematic from a weather point of view. That said, most cities had very little or no rain on the last three census days.

### Where were the commuter cyclists living and working?

Other posts on this blog have also covered some of these maps, but not for all cities.

Some of the following maps are animated to show both 2006 and 2011 results, and note that the colour scales are not the same for all maps. I’ve sometimes zoomed into inner city areas when these are the only places with significant cycling mode share. White sections on maps represent areas with low density, or where the number of overall commuters was very small (sorry I haven’t gone to the effort of making every map 100% consistent, but rest assured the areas in white are less interesting). Click on the maps to see more detail.

#### Canberra

Firstly home locations:

The cycling commuters mostly appear to be coming from the inner northern suburbs. I don’t know Canberra intimately, but Google maps doesn’t show a higher concentration of cycling infrastructure in this area compared to the rest of Canberra.

Secondly, bicycle mode share by work destination (at the larger SA2 geography):

The highest mode share was 12% in the SA2 of Acton, which is dominated by the Australian National University. Perhaps a lot of the university staff live in the inner northern suburbs of Canberra?

#### Melbourne

By home location:

Cycling mode share is highest for origins in the inner northern suburbs and has grown strongly since 2006. There’s also been some growth in the Maribyrnong  and Port Phillip council areas off a lower base.

By work location (note: this data is at the smaller destination zone geography):

Cycling to work boomed in inner Melbourne between 2006 and 2011, particularly to workplaces in the inner north. Princess Hill had the highest bike share of 14% in 2011 (possibly dominated by Princess Hill Secondary College employees), followed by a pocket of south-west Carlton that jumped from around 5% to 13%. Apart from the inner north, there were notable increases in Richmond, Balaclava, Yarraville and Southbank. Cycling rates within the CBD are relatively low, perhaps reflecting limited cycling infrastructure on CBD most streets in 2006 and 2011.

Firstly, by home:

Adelaide appears to lack any major concentrations of cycling, although slightly higher levels are found just outside the parkland surrounding the CBD.

Secondly, bicycle mode share by work destination at the (larger) SA2 geography:

The numbers are all small, with 3% in the (large) Adelaide CBD. I imagine a map based on destination zones might show some pockets with higher mode share, but that data isn’t freely available unfortunately.

#### Perth

By home location:

The inner northern and western suburbs, and south of Fremantle seem to be the main areas of cycling growth.

For workplaces at the larger SA2 geography:

The highest mode share was in ‘Swanbourne – Mount Claremont’, only slightly ahead of ‘Nedlands – Dalkeith – Crawley’ – which contains the University of Western Australia. The Fremantle SA2 (with 3% bicycle mode share by destination) includes of Rottnest Island where around 20% of the 73 resident commuters cycled to work, but the result will be easily dominated by the mainland Fremantle section.

#### Brisbane

By home location:

There was significant growth in cycling from the West End, and around the University of Queensland/St Lucia – which may be related to the opening of the Eleanor Schonell Bridge (after the 2006 census) which only carries pedestrians, cyclists and buses.

By work location (at larger SA2 geography):

The highest share was in St Lucia – which is probably dominated by the University of Queensland. Neighbouring Fairfield – Dutton Park came in second. These two areas are directly joined by the Eleanor Schonell Bridge which provides cycling a major advantage over private transport. It seems to have been quite successful at promoting cycling in these areas.

#### Sydney

First by home location:

There were quite noticeable shifts to cycling in the inner south and around Manly.

By work location (by smaller destination zone geography):

There was strong growth, again in the inner southern suburbs. In 2011 bicycle mode share was highest in Everleigh (11.5%) following by the University of NSW (Paddington) at 7.9% (excluding travel zones with less than 200 employees who travelled).

#### Rural Australia

Here’s a map showing bicycle share by SA2 workplace location for all of Australia, which gives a sense of bicycle mode shares in rural areas.

Higher regional/rural bicycle mode shares are evident in southern Northern Territory (Petermann – Simpson), Katherine (NT), the Exmouth region, the Otway SA2 on the Great Ocean Road in western Victoria, and Longford – Loch Sport in eastern Victoria. I’ll let other people explain those.

The SA2s in Australia with the highest cycling mode shares in 2011 (by home location) were:

• Lord Howe Island, NSW: 21%
• Acton, ACT (covering Australian National University): 12%
• Port Douglas, Queensland: 10%
• Parkville, Victoria (covering the University of Melbourne main campus): 8%
• East Side, Northern Territory (Alice Springs): 8%
• St Lucia, Queensland (covering the University of Queensland): 8%

### How far did people cycle to work? (in Melbourne)

It is difficult to get precise distances for journeys to work, but approximations are possible. I’ve calculated the approximate distance for each journey to work by measuring the straight line distance between the centroid of the home and work SA2s and then rounded to the nearest whole km. To give a feel for how this looks, here is a map showing inner Melbourne SA2s and the approximate distances between selected SA2s:

This distance measure generally works well in inner city areas. However in the outer suburbs SA2s are often much larger in size, and sometimes only partially urbanised. However as we’ve seen above the volumes of cycling journeys to work are very low in these places, so that hopefully won’t skew the results signficantly.

Two-thirds of cycling journeys to work in Melbourne were approximately 5km or less, with 80% less than 7 km, and 30% were 2 km or less.

The longest commute recorded within Greater Melbourne was approximately 44km.

### Was cycling combined with other modes?

The following chart shows that bicycles were seldom combined with other modes:

Around 16-17% of cycling commuters in the four largest cities in 2011 involved another mode. Use of other modes with cycling grew in all cities between 2006 and 2011

The next chart shows what these other modes were:

Sydney, Melbourne, Brisbane and Perth had high rates of bicycle use with trains, while combining car and bicycle was more common in the smaller cities.

The next chart shows the number of trips involving bicycle and trains in 2006 and 2011:

The chart shows the relative success of Melbourne Parkiteer program of introducing high quality bicycle cages at train stations, which has helped boost the number of people access the train network by bicycle by around 600 between 2006 and 2011. I understand a similar project has been undertaken in Perth which saw growth of around 250.

In Melbourne, the home locations for people using bicycle and train are extremely scattered – the following map shows a seemingly random smattering:

### How does commuter cycling vary by age and sex?

This chart shows remarkably clear patterns. Males were much more likely to cycle to work. Teenage boys (particularly those under driving age) had the highest cycling mode shares (with teenage girls much less likely to cycle). The next peak for men was around the mid thirties, and women’s mode share peaked around ages 28-32.

### Where are women more likely to cycle to work?

Women are sometimes talked about as the “indicator species” for cycling – ie if you have large numbers of women cycling compared to men then maybe you have good cycling infrastructure that attracts a broader range of people.

The census data can shed some light on this. For each SA2 in Melbourne, I have calculated the male and female cycling mode shares both as a home origin, and as a work destination (this analysis looks at people who only used bicycle (and walking) in their journey to work). I’ve then calculated the ratio of male mode share to female mode for each area (SA2).

I’ve used the ratio of mode shares in preference to the straight gender split of cycling commuters – as female workforce participation is generally lower and there can be spatial variations in the gender split of the workforce. 46% of all journeys within Greater Melbourne in the 2011 census were by females, but only 28% of cycling journeys to work were by females.

The following map shows the ratio of male to female cycling mode shares by home location for SA2s (with more than 50 commuter cyclists, and where the bicycle mode share is above 1%):

Areas attracting comparable female and male bicycle shares include the inner northern suburbs and – curiously – Toorak (probably many using the off-road Gardiners Creek and Yarra Trails to access the city centre).

Here’s a similar map, but by workplace areas:

The patterns are much more pronounced. Six SA2s had higher female mode shares than male: Yarraville, Fitzroy North, Brunswick East, Ascot Vale, Carlton North – Princes Hill, and Elsternwick.

The areas with near-1 ratios of male to female mode shares were similar to the areas with higher total cycling mode shares. The following chart confirms this relationship (note areas with cycling mode shares below 1% not shown):

What this also shows is that home-area mode shares reach much higher values than workplace-area mode shares. Perhaps the secret is in the home-area cycling infrastructure? Or perhaps it’s more to do with the residential demographics?

See the Bicycle Network Victoria website for more data about female cycling rates in Melbourne.

### Do women cycle the same distances as men?

Again using the approximate straight line commuting distances (as explained above) the following chart shows that women’s cycling commutes are a little shorter than men’s, but not by much:

The median female cycling commute was approximately 1.8 km shorter than for males.

### What types of workers are more likely to cycle to work?

Firstly, I’ve looked at the differences between public and private sector employees.

Before I dive into the data, it’s important to recognise that different types of workers are not evenly spread across Australia. Some types of jobs concentrate in city centres while others might be more likely to be found in the suburbs or the country. Therefore many of the following charts show results for Australia as a whole, but also for people working in central Melbourne (the SA2s of Melbourne, Carlton, Docklands, East Melbourne, North Melbourne and Southbank), which has a relatively high rate of cycling to work.

The data suggests public servants were much more likely to cycle to work:

The local government result has prompted me to calculate the cycling mode shares for local government workers across Australia (assuming workers work within the council for which they work). Here are bicycle mode shares for the top 20 councils for employee cycling mode share in the census:

 Council State Bicycle mode share Tumby Bay (DC) SA 23.5% Kent (S) WA 18.8% Carnamah (S) WA 16.0% Central Highlands (M) Qld 14.3% Uralla (A) NSW 13.8% Wakefield (DC) SA 13.5% Nannup (S) WA 12.5% Broome (S) WA 12.1% Alice Springs (T) NT 11.8% Narembeen (S) WA 11.5% Blackall Tambo (R) Qld 11.3% Kowanyama (S) Qld 11.2% Exmouth (S) WA 11.1% Yarra (C) Vic 10.4% Glamorgan/Spring Bay (M) Tas 8.7% Torres (S) Tas 8.6% Yarriambiack (S) Qld 8.3% Mallala (DC) Vic 8.0% Richmond Valley (A) NSW 7.2% McKinlay (S) Qld 6.7%

Most of the top 20 are non-metropolitan councils. Melbourne’s City of Yarra is the top metropolitan city council (within Greater Melbourne the next highest councils are Moreland 6.1%, Port Phillip 5.6%, Melbourne 5.6% and then Stonnington 4.9%).

National government employees had the highest bicycle mode share of all of Australia. I suspect this relates to university staff, as many of the earlier maps showed university campuses often had relatively high rates of employees cycling (85% of “higher education” employees count as “national government” employees).

The census data can also be disaggregated by income:

Cycling mode shares were highest for people on high incomes. Initially I thought this might reflect the fact that high income jobs are often in city centres were cycling is relatively competitive with private and public transport. However, even within central Melbourne workers, cycling rates are higher for those on high incomes (curiously with a second peak for those on incomes between \$300 and \$399 per week).

Does cycling to work make you healthier and therefore more likely to get promoted and earn a higher income? Or are employers offering workplace cycling facilities to attract highly paid staff? I haven’t got data that answer those questions.

Consistent with higher rates of cycling for higher income earners, those in more highly skilled occupations were more likely to cycle to work:

I suspect this might reflect the presence/absence of workplace cycling facilities (perhaps office workplaces are more likely to provide cycling facilities than retailers for example) and/or the ability to afford to live close to work (which makes cycling easier).

### Are recent immigrants more likely to ride to work?

This one really surprised me and I only investigated it because it was possible to do. The census asks what year people migrated to Australia (if not born here), and it turns out that recent immigrants were much more likely to cycle to work:

This might be explained by the demographics of recent immigrants (eg car ownership, home location, income levels, occupation and age).

I’d welcome comments on any other trends people might spot in the data.

## Are Australian cities becoming denser?

Tue 5 November, 2013

[Updated and fully revised June 2015 with June 2014 population data and 2011 density calculations using square kilometre grid population data. First published November 2013]

While Australian cities have been growing outwards with new suburbia, they have also been getting denser in established areas, and the new areas on the fringe are often more dense than growth areas used to be (see last post). So what’s the net effect – are Australian cities getting more or less dense?

This post also explores measures of population-weighted density for Australian cities large and small over time. It also tries to resolve some of the issues in the calculation methodology by using square kilometre geometry, looks at longer term trends for Australian cities, and then compares multiple density measures for Melbourne over time.

### Measuring density

Under the traditional measure of density, you’d simply divide the population of a city by the metropolitan area’s area (in hectares). As the boundary of the metropolitan areas seldom change, the average density would simply increase in line with population with this measure. But that density value would also be way below the density at which the average resident lives because of the inclusion of vast swaths of unpopulated land within “metropolitan areas”, and so be not very meaningful.

Enter population-weighted density (which I’ve looked at previously here and here). Population-weighted density takes a weighted average of the density of all parcels of land that make up a city, with each parcel weighted by its population. One way to think about it is the residential density in which the “average resident” lives.

So the large low-density parcels of rural land outside the urbanised area but inside the “metropolitan area” count very little in the weighted average because of their small population relative to the urbanised areas. This means population-weighted density goes a long way to overcoming having to worry about the boundaries of the “urban area” of a city. Indeed, in a previous post I found that removing low density parcels of land had very little impact on calculations of population-weighted density for Australian cities. However, the size of the parcels of land used in a population-weighted density calculation will have an impact, as we will see shortly.

Calculations of population-weighted density can answer the question about whether the “average density” of a city has been increasing or decreasing. But as we will see below, using geographic regions put together by statisticians based on historical boundaries is not always a fair way to compare different cities.

### Population-weighted density of Australian cities over time

Firstly, here is a look at population-weighted density of the five largest Australian cities (as defined by ABS Significant Urban Areas), measured at SA2 level (the smallest geography for which there exists a good consistent set of time-series estimates). SA2s roughly equate to suburbs.

According to this data, most cities bottomed out in density in the mid 1990s. Sydney, Melbourne and Perth have shown the fastest rates of densification in the last three years.

What about smaller Australian cities? (120,000+ residents in 2014):

Darwin comes out as the third most dense city in Australia on this measure, with Perth rising quickly in recent years to be equal to Brisbane. Most cities have shown densification in recent times, with the exceptions being Geelong, Hobart, and Townsville.

However, we need to sanity test these values. Old-school suburban areas of Australian cities typically have a density of around 15 persons per hectare, so the values for Geelong, Newcastle, Darwin, Townsville, and Hobart all seem a bit too low for anyone who has visited them. I’d suggest the results may well be an artefact of the arbitrary geographic boundaries used – and this effect would be greater for smaller cities because they would have more SA2s on the interface between urban and rural areas (indeed all of those cities are less than 210,000 in population).

For reference, here are the June 2014 populations of all the above cities:

The following map shows Hobart, with meshblock boundaries in black (very small blocks indicate urban areas), SA2s in pink, and the Significant Urban Area (SUA) boundary in green.  You can see that many of the SA2s within the Hobart SAU have pockets of dense urban settlement, together with large areas that are non-urban – ie SA2s on the urban/rural interface. The density of these pockets will be washed out because of the size of the SA2s.

### Reducing the impact of arbitrary geographic boundaries

As we saw above, the population-weighted density results for smaller cities were very low, and probably not reflective of the actual typical densities, which might be caused by arbitrary geographic boundaries.

Thankfully ABS have followed Europe and released of a square kilometre grid density for Australia which ensures that geographic zones are all the same size. While it is still somewhat arbitrary where exactly this grid falls on any given city, it is arguably less arbitrary than geographic zones that follow traditional notions of area boundaries.

Using that data, I’ve been able to calculate population weighted density for the larger cities of Australia. The following chart shows those values compared to values calculated on SA2 geography:

You’ll see that the five smaller cities (Newcastle, Hobart, Geelong, Townsville and Cairns) that had very low results at SA2 level get more realistic values on the kilometre grid.

You’ll notice that most cities (except big Melbourne and Sydney) are in the 15 to 18 persons per hectare range, which is around typical Australian suburban density.

While the Hobart figure is higher using the grid geography, it’s still quite low (indeed the lowest of all the cities). You’ll notice on the map above that urban Hobart hugs the quite wide and windy Derwent River, and as such a larger portion of Hobart’s grid squares are likely to contain both urban and water portions – with the water portions washing out the density (pardon the pun!). While most other cities also have some coastline, much more of Hobart’s urban settlement is near to a coastline.

But stepping back, every city has urban/rural and/or urban/water boundaries and the boundary has to be drawn somewhere. So smaller cities are always going to have a higher proportion of their land parcels being on the interface – and this is even more the case if you are using larger parcel sizes. There is also the issue of what “satellite” urban settlements to include within a city which ultimately becomes arbitrary at some point. Perhaps there is some way of adjusting for this interface effect depending on the size of the city, but I’m not going to attempt to resolve it in this post.

### International comparisons of population-weighted density

So now that I have calculated population weighted density of Australian cities using a kilometre grid – I invite other analysts to do the same calculations for other cities of the world – and then we might have a much fairer comparison of city densities (although still not perfect).

### Changes in density of larger Australian cities since 1981

We can also calculate population-weighted density back to 1981 using the larger SA3 geography. An SA3 is roughly similar to a local government area (in Melbourne at least), so getting quite large and including more non-urban land. Also, as Significant Urban Areas are defined only at the SA2 level, I need to resort to Greater Capital City Statistical Areas for the next chart:

This shows that most cities were getting less dense in the 1980s (Melbourne quite dramatically), with the notable exception of Perth. I expect these trends could be related to changes in housing/planning policy over time. This calculation has Adelaide ahead of the other smaller cities – which is different ordering to the SA2 calculations above.

When measured at SA2 level, the four smaller cities had almost the same density in 2011, but at SA3 level, there is more separating them. My guess is that the arbitrary nature of geographic boundaries is having an impact here. Also, the share of SA3s in a city that are on the urban/rural interface is likely to be higher, which again will have more impact for smaller cities. Indeed the trend for the ACT at SA3 level is very different to Canberra at SA2 level.

### Melbourne’s population-weighted density over time

I’ve taken a more detailed look at my home city Melbourne, using all available ABS population figures for the geographic units ranging from mesh blocks to SA3s inside “Greater Melbourne” (as defined in 2011) or inside the Melbourne Significant Urban Area (SUA, where marked), to produce the following chart:

Note: I’ve calculated population-weighted density at the SA2 level for both the Greater Capital City Statistical Area (ie “Greater Melbourne”, which includes Bacchus Marsh, Gisborne and Wallan) and the Melbourne Significant Urban Area (slightly smaller), which yield slightly different values.

All of the time series data suggests 1994 was the turning point in Melbourne where the population-weighted density started increasing (not that 1994 was a particularly momentous year – the population-weighted density increased by a whopping 0.0559 persons per hectare in the year to June 1995 (measured at SA2 level for Greater Melbourne)).

You’ll also note that the density values are very different when measured on different geographic units. That’s because larger units include more of a mix of residential and non-residential land. The highest density values are calculated using mesh blocks (MB), which often separate out even small pockets of non-residential land (eg local parks). Indeed 25% of mesh blocks in Australia had zero population, while only 2% of SA1s had zero population (at the 2011 census). At the other end of the scale, SA3s are roughly the size of local councils and include parklands, employment land, rural land, airports, freeways, etc which dilutes their average density.

In the case of SA2 and SA3 units, the same geographic areas have been used in the data for all years. On the other hand, Census Collector Districts (CD) often changed between each five-yearly census, but I am assuming the guidelines for their creation would not have changed significantly.

Now why is a transport blog so interested in density again? There is a suggested relationship between (potential) public transport efficiency and urban density – ie there will be more potential customers per route kilometre in a denser area. In reality longer distance public transport services are going to be mostly serving the larger urban blob that is a city – and these vehicles need to pass large parklands, industrial areas, water bodies, etc to connect urban origins and destinations. The relevant density measure to consider for such services might best be based on larger geographic areas – eg SA3. Buses are more likely to be serving only urbanised areas, and so are perhaps more dependent on residential density – best calculated on a smaller geographic scale, probably km grid (somewhere between SA1 and SA2).

## The growth of Melbourne 1986-2011, animated

Tue 29 October, 2013

Following on from my recent post about the changing socio-economic landscape of Melbourne, this post simply looks at the changing shape and density of urban Melbourne using 5-yearly census data at collector district (1986-2006) and SA1 level (2011).

Straight to it: here is map of Melbourne residential density, click to enlarge and animate:

You can see the sprawl of Melbourne over the years, including changes that suggest shifts in the urban growth boundary after development previously seemed to have stopped against a line (particularly evident on the western edge of the City of Brimbank).

Here is another animated map showing the inner city area, with a density scale ranging from 10 to 100 persons/ha, so you can distinguish higher densities than the map above. Click to enlarge and animate.

You can see a lot more going on in established areas on this map, including densification in the CBD, St Kilda, St Kilda Road (conversion from office space), Parkville, Port Melbourne around Bay Street, Kensington Banks, Brunswick, Fitzroy, Southbank, South Melbourne, Elwood, Maribyrnong, Carlton, and many more.

A few things to note:

• The size of the districts changes each year, particularly around the fringe. You’ll often see a large red patch where a larger block is only partly inhabited in one year, only to be replaced by smaller denser patches in future years. Patches of green that disappear might be the enlargement of a district causing a blending out of a small pocket of high density, rather than an actual drop in density.
• Shades of pink indicate densities between 5 and 10 per hectare on the large map, and between 10 and 20 per hectare on the inner map. Lower densities are shown as white.
• In 2011 the ABS changed their statistical geography. I have used SA1s from 2011 as the most comparable area unit to a census collector district, however they are generally smaller and so densities may appear to jump slightly in 2011 in some areas.

## Visualising the changing socio-economic landscape of Melbourne

Sun 29 September, 2013

This post is drifting a little away from transport, but I hope you will find this interesting…

How has the spatial distribution of socio-economic advantage and disadvantage changed over time in Melbourne? (oh, and Geelong too)

The animated maps below are fascinating, but of course there’s lots of important caveats regarding the data.

Since 1986, the Australian Bureau of Statistics (ABS) has calculated Socio-Economic Indexes For Areas (SEIFA) based on five-yearly census data. These include indexes of relative socio-economic disadvantage (IRSD), and – since 2001 – an index of relative socio-economic advantage and disadvantage (IRSAD). For 2006 and 2011, SEIFA was explicitly designed to measurepeople’s access to material and social resources, and their ability to participate in society” (with similar intent for prior years).

This post looks at the spatial changes over time in these index values. I must be upfront: ABS explicitly cautions this type of analysis. This is mostly because the component census variables that make up SEIFA scores and their respective weightings vary between each census, but also because statistical area boundaries change, the number of areas has increased, and indexes were calculated on usual residents from 2006 onwards (as opposed to people present on census night for 2001 and earlier). ABS also notes that middle range scores are very similar, so time-series analysis should focus more on the top and bottom ends of the spectrum. More discussion on this issue is available from ABS and .id consulting.

However, I’m going ahead noting the above (as readers also should!), on the following basis:

• The intent of the indexes has not changed over time, although the quality has (perhaps one day ABS will recalculate SEIFA values for previous census using better measures where possible)
• I’ve used percentile ranks within Victoria to get around the issue of the changing meaning of particular index values (although this might cause some issues if there has been a relative difference in changes between Melbourne and regional Victoria)
• I’ve included a summary of the component variables that have changed between censuses (documentation is available from 1996 onwards)
• I’m mapping this at a metropolitan scale with a view to looking at regional variations, rather than very local changes. In the following maps you’ll see fairly strong regional patterns
• My analysis will focus only on substantial shifts (which have indeed occurred)
• Excessive caution may mean that we never do any interesting analysis!

### Changes in Index of Relative Socio-economic Disadvantage (IRSD)

This index has been available from 1986 onwards.

More significant changes in the make up of this index in recent years include:

• 2011 added: families with jobless parents
• 2011 dropped: indigenous persons, renting housing from government authority
• 2006 added: household overcrowding (replacing multiple-family households), low rent payments, lack of an internet connection, low skill community and personal services workers, people who need assistance with core activities
• 2006 dropped Elementary Clerical, Sales and Service workers and tradepersons
• 2006 changed the evaluation of household income to consider ‘equivalised household income’ replacing a number of measures that try to capture income levels relating different household make-up scenarios. It also stopped using gender specific measures of people with certain occupations or unemployed
• 2001 saw no changes to the included variables from 1996
• Variables for persons who did or didn’t finish year 12 at school have changed slightly in both 2006 and 2011

check the SEIFA documentation for full details.

Click on this map to enlarge and see an animation of IRSD percentile values for the years 1986 to 2011.

You can see some quite dramatic changes over time. Two big trends of note are:

• Most inner city suburbs have gone from being some of the most disadvantaged to much less disadvantaged. It’s hard to imagine suburbs such as South Yarra and East Melbourne as being highly disadvantaged, but the data suggests that was the case in the 1980s. During this transformation, pockets of high disadvantage have remained, probably reflecting older government housing estates. There appears to have a been a fairly large change between 1986 and 1991. This could represent dramatic demographic change and/or reflect changes in the calculations of SEIFA index values.
• Areas with the highest disadvantage have generally shifted away from the city centre (including some middle suburbs such as Carnegie), perhaps reflecting the growth in high-end CBD jobs driving the attractiveness of near city living.
• New urban fringe growth areas often begin with low levels of disadvantage, but have become more disadvantaged over time. This is particularly evident in areas such as Hoppers Crossing, Werribee, Melton, Deer Park, Craigieburn, Keysborough, Karingal, Epping, Hampton Park, Cranbourne, Altona Meadows and Keilor Downs. Perhaps this is because when these areas were initially settled there were many double-income-no-kids households that now have more kids and less income? It could also be a reflection of a turnover in the resident population.
• The maps only show geographic units with a population density of 5 per hectare or more, so you can also see the urban growth of Melbourne (more on that in a upcoming post).

This index was first calculated in 2001 and aims to also measure advantage, not just factors that suggest disadvantage. In 2011 it included all but one of the IRSD variables, plus a number that describe levels of advantage (eg high income, higher education, occupations such as managers and professionals, high rent or mortgage payments, spare bedrooms).

The component variables of IRSAD have changed in line with the changes to IRSD, plus some other variables:

• 2011 added people with occupation classed as managers, houses with spare bedrooms, households with 3 or more cars
• 2006 added people paying low/high rent, high mortgage payments, renting from government authority, households with no car, households with broadband internet connection (replacing persons using the internet at home)

Again, check the SEIFA documentation for full details.

An aside: SEIFA associates higher car ownership with advantage, but I suspect some inner-city types might consider not needing to own a car an advantage.

Here is an animation of the Index of Relative Socio-economic Advantage and Disadvantage for years 2001 to 2011.  Again, click to enlarge and see the animation.

The changes between 2001 and 2011 are much less dramatic, probably because of the shorter time span. Some observations:

• The Melbourne CBD drops in 2011 – possibly because of a change of demographics (more students?) and/or a change in the component variables.
• Many parts of the middle eastern suburbs (particularly the Whitehorse area) appear to drop from the upper to the middle percentiles in 2011.

What’s also interesting to see is some socio-economic fault lines in Melbourne, such as:

• Altona North versus South Kingsville/Newport (north-south divide along Blenheim Road/Hansen Street/New Street)
• Skeleton Creek between Point Cook (including Sanctuary Lakes) and Altona Meadows
• A north-south line being the boundary between the Shire of Melton and the City of Brimbank in the north-western suburbs
• Along Hume Drive / Lady Nelson Way (an east/west line in northern Brimbank)
• Greenvale versus Meadow Heights (split by the proposed north-south Aitken Boulevard)
• A north-south divide through Heidelberg Heights, roughly parallel to the Hurstbridge rail line
• Along the Dingley Arterial between Dingley Village and Springvale

### How different are IRSAD and IRSD values?

IRSAD contains a lot more variables and uses different weightings. See the ABS website for full details.

For those who are interested in the correlation between the two, here’s a scatter plot for both 2006 and 2011 data comparing the two index values (as percentile ranks) for all CDs and SA1s (respectively) in Victoria:

You can see the relationships between the two indexes is stronger in 2011 (R-squared = 0.96) versus 2006 (R-squared = 0.89). This might reflect the make up of the variables in each year and/or the smaller geographic units in 2011 (SA1s) which may reduce diversity within each geographic unit.

I’m sure others could spot other interesting patterns, and/or offer explanations for the changes over time (comments welcome).