refine: Clean up raw records
In owenjonesuob/BANEScarparking: Car Parking Data for Bath and North East Somerset

Description Usage Arguments Value Examples

Uses functions from the dplyr package to clean up raw records obtained from the Bath: Hacked datastore. The process is as follows:

Select columns containing useful information only
Remove any records with NA entries
Remove records for "test car park"
Convert Name and Status to factors
Remove records with negative occupancies
Calculate Proportion column (Occupancy/Capacity)
Remove records with Proportion greater than max_prop
Remove duplicate records (see first_upload)

1	refine(x, max_prop = 1.1, first_upload = FALSE)

`x`	A data frame containing records to be cleaned up (e.g. the data frame obtained by calling `get_all_crude`).
`max_prop`	The point at which records are discarded due to overly-full Occupancy values (default is 1.1, or 110% full, to allow for circulating cars).
`first_upload`	If `TRUE`, ensures that when duplicate records are removed, the copy which is kept is the first one uploaded after the record was taken. This takes much longer to run, due to sorting.

A data frame of clean records, with 7 columns:

Name: The name of the car park where the record was taken.
LastUpdate: The time the record was taken (POSIXct date-time object).
DateUploaded: The time the record was uploaded to the Bath: Hacked database (POSIXct date-time object).
Occupancy: The total number of cars in the car park.
Capacity: The number of parking spaces in the car park.
Status: Description of the change in occupancy since the previous record from that car park.
Proportion: Calculated as (Occupancy/Capacity).

## Not run: 
raw_data <- get_all_crude()
some_records <- raw_data[1:1000, ]

dim(some_records)
## 1000   16

df <- refine(raw_data)
dim(df)
## 813   7

## End(Not run)