GTFS filter trips in a given date

[temospena]

The function tidytransit::filter_feed_by_date() retrieves all the services that run in that date but that also sun in other dates.
This is not what we want.

Another option is to identify which service pattern has more days, trips and routes, and filter that service_id.

Here is how:
https://cran.r-project.org/web/packages/tidytransit/vignettes/frequency.html

[valenca13]

Here are the Results for Braga. Braga has a total of 11 service patterns. It is important to mention that the Calendar dates are from: 28/03/2024 - 04/04/2024.

[valenca13]

AML has 188 service patterns. We can choose the ones that operate during more days the year. Calendar dates from AML range from 01/07/2023 to 31/12/2024.

[temospena]

Can you choose more than one?

…

On Tue, Apr 9, 2024 at 5:07 PM Gabriel Valença ***@***.***> wrote: lisbon.png (view on web) <https://github.com/U-Shift/SiteSelection/assets/69808511/ae35a94b-5d5b-433c-8570-ec34001ffad9> AML has 188 service patterns. We can choose the ones that operate during more days the year. Calendar dates from AML range from 01/07/2023 to 31/12/2024. — Reply to this email directly, view it on GitHub <#9 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AJKLUXWFQ6BL36653XKJ75LY4QG3XAVCNFSM6AAAAABFZMNIXCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDANBVGU3DOOJQGA> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[valenca13]

Analyze the service patterns of AML. Take a look at the locations.

[valenca13]

All the bus stops that are present in the AML. Check the locations of the most frequent service patterns. Are they in different locations?

[valenca13]

Bus routes of service pattern "s_d38ffee" (192 days)

[valenca13]

Bus routes of service pattern "s_0973a74" (191 days)

[valenca13]

Bus routes of service patterns "s_70dfe23" (188 days)

[valenca13]

The most frequent service patterns are actually in different locations in the AML. Should we aggregate the 3? What are the next steps? @temospena

[temospena]

I would expect that. But there should be 4 expressive service patterns, not 3, 1 for each area.
After finding this 4th one, I would suggest to change the calendar dates of the less expressive ones, to be compatible with the most expressive one.
For instance, if service_pattern 3 is running between 01-01-2024 to 31-03-2024 and service_pattern 1 is running between 01-10-2023 to 31-12-2023, recode the calendar dates of this one to be at same date and time than the previous one (looking at the calendar and matching monday with monday)

[valenca13]

This is the service pattern with that operates in 7th most days of the year ("s_a6bf59d (44 days)). The other 3 service patterns are very similar to the ones that I have already got.

[valenca13]

Service pattern 10 ("s_ba5dcc4" (39 days)). The routes appear more in the north with 143 routes (ranked 26th in terms of number of routes)

[valenca13]

Service pattern 11 seems to be the fourth region ("s_f22b75f" (39 days)). @temospena

[valenca13]

Also service pattern 12 (""s_fff1bcb" (39 days)) @temospena

[temospena]

Service pattern 11 seems to be the fourth region ("s_f22b75f" (39 days)). @temospena

There should be a service pattern going south, to Montijo, Alcochete, and so on...

[valenca13]

Service pattern 11 seems to be the fourth region ("s_f22b75f" (39 days)). @temospena

There should be a service pattern going south, to Montijo, Alcochete, and so on...

Service_pattern "s_bc376dc" (37 days)

[valenca13]

Adopted service patterns (ranked by "days in service"):

1. Service pattern 1: "s_d38ffee" (192 days)
2. Service pattern 2: "s_0973a74" (191 days)
3. Service pattern 3: "s_70dfe23" (188 days)
4. Service pattern 12: "s_fff1bcb" (39 days)
5. Service pattern 18: "s_bc376dc" (37 days)

[valenca13]

5. "s_bc376dc"

O service pattern 18 só opera 1 mês. Vou ter que analisar os próximos passos. @temospena

[valenca13]

• Filtrar por essa data (e.g., 10 de abril 2024) para cada servicepattern_id.
• Calcular o numero de frequencias por cada paragem depois juntar todas tabelas em uma só.

[valenca13]

• Lisboa está dando erro. Não acha dados.
• Juntar dados com rbind
• Fazer o piggyback para colocar os dados (geopackage) no release

[valenca13]

Critérios:

• Classificar a frequência por hora nos níveis de serviço (1-4), baseados nos critérios da foto.
• Calcular o nível de serviço máximo ao decorrer de todas as horas.
• Critério de inclusão: Nivel de serviço máximo 3 ou 4.

@temospena

[valenca13]

@temospena Já fiz o código dos níveis de serviço dentro da célula no código "cellcandidates.R". Por favor verifique se está tudo ok. Obrigado.

[temospena]

The file was generated for Portugal (great work @valenca13 !) and is available at releases.

The code is not fully reproducible because tidytransit](tidytransit::set_servicepattern() generates random hash numbers for identificators.

Something to consider when rebuilding for other year/region.