add_distance() calculates the distance between sensors based on a given corridor and direction. Interpolates missing values by default.
generate_spatial_lines() generates an sf object with a LINESTRING for each given corridor, corridor direction, and corridor category. Limitations are noted in documentation.
Function update aggregate_sensor_data() is now aggregate_sensor()
_data() is superfluous, so was removed.
New name reinforces the function's purpose, which is to aggregate data from a single sensor.
New logical paramenters
replace_impossible, replace impossible volume and occupancy values in the 30-second (raw) data with NA
interpolate_missing, interpolate missing volume and occupancy using a rolling mean of the two neighboring observations (before and after).
occupancy_threshold, set a threshold for occupancy when calculating speed. Very low occupancy percentages were causing
Made significant progress on "Calculate speed and delay" vignette
Chose a subset of sensors to consistently work with
Used {mice} to impute null speed values with a random forest model based on total volume, total occupancy, hour, and day type.
Added bibliography with regional and international relevant literature
New plots
Plot proportion of all speed values that are NA by hour for 10-minute and 1-hour aggregations.
Plot relationship between speed, occupancy % and relative volume for 10-minute and 1-hour aggregations.
Plot observed and imputed speed density at 1-hour aggregations by day type.
Plot imputed speed values at 1-hour aggregation by hour and day type.
Plot speed and delay using the median speed from hours 1-5 and 20-24 (chosen arbitrarily) by hour.
General updates
Shift from using {dplyr} to {data.table}
Added a NEWS.md file to track changes to the package.
tc.sensors 0.2.0
add_distance()calculates the distance between sensors based on a given corridor and direction. Interpolates missing values by default.generate_spatial_lines()generates ansfobject with aLINESTRINGfor each given corridor, corridor direction, and corridor category. Limitations are noted in documentation.aggregate_sensor_data()is nowaggregate_sensor()_data()is superfluous, so was removed.replace_impossible, replace impossible volume and occupancy values in the 30-second (raw) data withNAinterpolate_missing, interpolate missing volume and occupancy using a rolling mean of the two neighboring observations (before and after).occupancy_threshold, set a threshold for occupancy when calculating speed. Very low occupancy percentages were causing{mice}to impute null speed values with a random forest model based on total volume, total occupancy, hour, and day type.NAby hour for 10-minute and 1-hour aggregations.{dplyr}to{data.table}NEWS.mdfile to track changes to the package.This PR closes #11