Calculate travel demand (actual travel patterns) between zones

[Hussein-Mahfouz]

Dataset

I need a dataset with actual travel patterns (ideally it would have different purposes and be disaggregated by time of day)

Options include:

1. The Synthetic Population Catalyst dataset (see notes here: https://github.com/Hussein-Mahfouz/Lit-Review/issues/10#issuecomment-1592998427)
2. The United Kingdom Time Use Survey, 2014-2015:
   • This is for a sample of the population only (sample size?).
   • Does the data have a geographical component? Is it point locations or are activities assigned to an OA?
   • Given spatial component, the data could still be used to get relative importance of different OD pairs
3. Create an activity-based dataset.

• Examples
  • A dynamic microsimulation model for epidemics
  • Synthetic population and travel demand for Paris and Île-de-France based on open and publicly available data
• Relevant issue: Create activity-based model  #19

Trip purposes

What trip purposes should be considered for the MCA measure? (#5). Maybe this should be based on the trip purposes in existing datasets / studies:

• A dynamic microsimulation model for epidemics: "Currently three activities, other than spending time at home, are simulated in the model: working, attending school and shopping".
• Trip purposes in SPC: shopping/retail, primary school, secondary school, working, staying at home - relevant code here

[Hussein-Mahfouz]

The census 2021 flow data is out [blogpost link]. The origin-destination data can be downloaded from nomisweb.

• It is available at MSOA and OA level: ODWP03EW
• You can also disaggregate the data by different variables, including car or van availability: ODWP14EW

[Hussein-Mahfouz]

The census 2021 data is not disaggregated by method of travel. The 2011 data was disaggregated - see table WU03EW

I emailed census.customerservices@ons.gov.uk and they said that I have to fill out a specific request form to request a dataset (see link here). They will begin the request processing service in January 2024, but the requests are not free. This is the breakdown:

[Robinlovelace]

Happy to support this with my personal development account if needs be Hussein, great bit of research.

[Hussein-Mahfouz]

Response from ONS (check text in bold). I don't understand how this work would take a month to compile. What do they mean by "tabulate"? If the data exists (like the 2011 census) then it seems like a simple sql query to me.

If the data is not available in the ready-made multivariate datasets or using the create a custom dataset tool, you may be able to commission a table to your own specifications.

However, Origin Destination tables can contain large amounts of data so if you were to commission a dataset it is worth noting the following points:

Generally, customers don’t request OD workplace at MSOA level. Customers tend to select an area of residence by all the MSOAs of workplace or all the areas of residence by a selected area of workplace.

This is because we have 7264 MSOAs in England and Wales as area of residence. We have 7264 MSOAs of workplace + categories ‘work in Scotland or Northern Ireland’, ‘Work on an offshore installation’, ‘work outside the UK, ‘No fixed place of work’, so a total 7268 areas.

If you multiply 7264 x 7268 = 52,794,752 combinations to cross tabulate just for the geography alone. When this is multiplied by the variable needed, in this case ‘Method of travel to work’ (5 categories for example) 52,794,752 x 5 = 263,973,760 rows.

This is not a manageable file. We can truncate it, but will end up with around 25 to 30 files. The cost would be around £350+VAT for the first file, £175 + VAT for all the remaining files. This represents around a month of work or more. Not all cross tabulations will have counts in, but we still have to tabulate them.

My response

Nomisweb has a selection query for census 2011 which allows me to get MSOA-to-MSOA level flows by method of travel to work - table WU03EW (link to the page). The query is capped at 5,000,000 cells. Is there a similar service for census 2021 where I can query MSOA level OD data by method of travel (and stick to a given cap)?
I have tried the 'create a custom dataset' tool that you mention but i cannot recreate the query from the 2011 census

For context, I am trying to get MSOA-level OD data by method of travel to work for Leeds (or West Yorkshire if the cap permits)

ONS

Thank you for your follow up email, I’m glad the data I sent is helpful.

Nomis have origin destination datasets on their backlog but are not yet able to give any indication about release dates beyond confirming that it will not be imminent.

[Robinlovelace]

Keep up the good work Hussein and let me know if I can help in any way.

[Hussein-Mahfouz]

Commuting data by mode for census 2021 is available at WICID but you have to log in with a uni account. It's at MSOA level only.

File name is ODWP02EW_MSOA

Thanks for pointing me to this @Robinlovelace !

[Hussein-Mahfouz]

ODWP15EW_MSOA: "The estimates classify people currently resident in each Middle layer Super Output Area (MSOA), by their method of travel to work, and by sex, and by age. They also show the movement between their area of residence and workplace MSOA"

[Robinlovelace]

Crazy they didn't point you, or presumably others, to the free version at WICID no?

[Hussein-Mahfouz]

Yeah they basically said the dataset doesn't exist yet ...