search

atfutures-labs / ipums-timeuse

Analyse historical transit behaviour from IPUMS time use data
2 stars 0 forks source link

how to meaure modal split #1

Open mpadge opened 6 years ago

mpadge commented 6 years ago

@Robinlovelace This repo is intended to hold reproducible form of code used to generate this result: junk

Cycling values are multiplied by 5. This provides a useful estimate for typical historical rates of modal change. I've analysed times here because they're easier than journey numbers. The data contain historical records of 862,000 car journeys, 53,000 walking trips, and (unfortunately only) 2,885 bike trips. Each of these record time of day, duration, and (for most) trip purpose. I've just aggregated durations to generate this rough first go. Public transport data are also included, but not shown here.

I extracted data for all 8 countries with time use data, but could only obtain useful data for the US and the UK. Interesting that the cycling downtick in the UK post-2000 that James emphasised comes out quite clearly: 2001 relative travel time spent on bike = 2.56% (5 = 12.8), decrease by 2005 to 1.29% (5 = 6.4). (High rates of walking in the UK pre-1987 are anomalous.) Finally, note the interesting suggestion here that relative time spent cycling compared with walking is considerably less in the UK than in the US (although naturally more in absolute terms).

Can you now please advise on how best to quantify modal share?

mpadge commented 6 years ago

Note that there are also demographic stats such as age and gender, but there likely aren't enough data to accurately analyse these sub-groups

Robinlovelace commented 6 years ago

Interesting stuff, very interesting. Modal split: get it from a travel survey, e.g. NTS. One to discuss next week I think (sorry for taking > 7 days to respond!).