To do: Sao Paulo

[robj411]

This is a to-do list for the Sao Paulo setting, which includes data, parameters, and any special functions required. Please add to it any items that are missing, and tick off items when they are complete.

• [x] Trips data @usr110 to check

• [x] PA data (pa_sao_paulo.csv file with columns: "sex", "age", "work_ltpa_travel_marg_met") @usr110 to check

• [x] Population data (population_sao_paulo.csv file with columns: "sex", "age", "population"). The age boundaries here should align with the GBD boundaries, and they define the age groups on which the calculation is executed. Accra data is being used here @leandromtg

• [x] Emission inventory -- motorcycle=4, car=4, bus_driver=32, big_truck=56, truck=4, van=0, other=0, taxi=0. Table 5 of https://cetesb.sp.gov.br/veicular/wp-content/uploads/sites/6/2013/12/Plano_de_Controle_de_Poluicao_Veicular_do_Estado_de_Sao_Paulo_2014-2016.pdf

• [x] distribution for BUS_WALK_TIME, if walk to bus is added Don't add

• [x] distribution for PM_CONC_BASE (background PM2.5) -- mean: 20. Distribution: lnorm(log(20),log(1.35)) @robj411 to explore writing distributions in real terms-- mean and SD

• [x] distribution for PM_TRANS_SHARE (transport share of PM2.5) -- Mean: 0.4. Distribution: Beta(8,8) Not the real data, need to be explored in VOI, this data point for Sao Paulo was already in doubt

• [x] Fleet-to-motorcycle ratio, 175.48 / (154.08 + 175.48) if constant and ~ Beta (175.48 + 1 , 154.08 + 1 ) if uncertain -- see #63

• [ ] distribution for BACKGROUND_PA_SCALAR (amount to scale non-zero, non-travel PA)

• [ ] number for BACKGROUND_PA_CONFIDENCE (how confident are you about the number of zeros in the PA survey on a scale of 0 to 1?)

• [x] distribution for INJURY_REPORTING_RATE (rate at which we think injuries were reported) -- beta(50,3)

• [ ] distribution for CHRONIC_DISEASE_SCALAR (amount to scale NCDs) @leandromtg to do sao-paulo state vs country scaling factors for all-cause NCDs

• [x] number/distribution for DAY_TO_WEEK_TRAVEL_SCALAR -- 7

• [x] distribution for BUS_TO_PASSENGER_RATIO (if bus drivers are added) -- beta(20,600), 98.5% between 1/20 and 1/60 India's ratio for pass-km to busveh-km; using a distribution based on that; to be improved with more data; In future need to link this distribution to transport emissions, for now, they are independent

• [x] distribution for TRUCK_TO_CAR_RATIO (if truck drivers are added) -- Beta(3,10) Use a reasonable number for now and check how it impacts the results

[rahulatiitd]

@robj411 I am surprised to see so much missing from our very strong case study Sao Paulo. Can I just confirm, that when you say you need 'distribution' of something, you already have the mean value that you can work with.

[robj411]

Hi Rahul, I don't have these numbers, no, but if you do, you could add them to the above and mark it complete, e.g.

• [x] distribution for PM_TRANS_SHARE (transport share of PM2.5) -- mean: 0.4. Distribution: 95% within 0.2 and 0.6, OR Beta(8,3)

[leandromtg]

Sao Paulo Metropolitan Area population data added. Data from 2010 census adjusted for 2017 population projection.