bayareametro/clpr

R package with analysis utilities and approaches for a data set of obscured and anonymized Clipper smart card transactions.

clpr's Issues

put the vignette in the /vignettes folder

improve documentation of time change sql

https://github.com/BayAreaMetro/clpr/blob/master/inst/sql/day_fares.sql#L4

write what this line does in a comment above the query

write an exploratory document on what transactions constitute a transfer and in what domains (discounts, travel model, etc)

the lag method used here works for a specific desired outcome but it might be helpful to compare it against results from the transfer operator column tag.

https://github.com/BayAreaMetro/clpr/blob/master/R/transfers.R#L36-L46

one helpful thing to do would be to get better documentation on how the transfer_operator column works (how its assigned, rules about what constitutes a transfer, etc.)

add necessary library loads to main readme

we should add this to the top of the main readme:

library(DBI)
library(dbplyr)
library(dplyr)

@ytse17 if you want to add this as you work in it feel free otherwise ill add it after.

move drop_tagons work to database

in theory we can probably move this processing to the database side.

currently it happens on the client side and takes a long time.

https://github.com/BayAreaMetro/clpr/blob/master/R/complete_trip.R#L30-L36

create deduplicated product metadata table

see https://github.com/BayAreaMetro/DataServices/blob/master/Project-Documentation/clipper/products.md#notes

make function names generic/consistent

rename drop_tagons to as_rides
rename bart_transactions_as_transfers to as_bart_journeys

this will make things more comprehensible.

complete audit of importfrom and namespace

this is currently being done on the fix-namespace branch:

https://github.com/BayAreaMetro/clpr/tree/fix-namespace

fix examples to work with pkgdown

source keys within the example

enable compatibility with data store tables

remove deprecated sql

remove extract_ sql from here, since its not relevant to the database.

add a link in the readme to the commit where its removed.

https://github.com/BayAreaMetro/clpr/tree/master/inst/sql

specify a @returns value in the docs for create_transfers_df()

specifically describe columns that are returned.

better documentation on using bart_transactions_as_transfers function

https://github.com/BayAreaMetro/clpr/blob/master/R/transfers.R#L84-L102

the vignette is helpful for showing how this function is used with the database connection, but needs documentation explaining how to use this function on dataframes extracted from the db

move the documentation to dataservices

https://github.com/BayAreaMetro/DataServices/tree/master/Project-Documentation/clipper

isolate the pull_sample_day in vignette to call it out

text before code, describing what it is. important to highlight it

either clarify how to use legacy functions or remove them

some of them do require box access and this may not be clear in the function docs. link to vignettes might help. some of them may also be redundant/less useful than functions that relate to data lake.

use summarise instead of transmute to count transactions per user per product type

should be better syntactically

add links to data services documentation as dependency to readme

move travel model examples to their own vignette

show return values in pkgdown site

use @ return instead of @ returns

Data availability?

It would be interesting if the public could also play with this data, even if it is a few years out of date. Is the data available as a dump anywhere online for public consumption?

fix links in the readme to work with pkgdown

update docs on products based on conversation with jen

document product names (e.g. at least say NA=Cash)

can do this in the function help docs @ returns

remove knitrr::kable call - just use head()

build a function that can be applied, for example in mutate, to return a vector of tagged journeys based on transactions

related to #14

recursively apply a set of rules (time between tags, transfer_operator value, etc) between individual transactions to determine their relationship.