Skip to contents

Classify Time Series Data

Source: R/classify-time-series-data.R
classify_time_series_data.Rd

Time series data will be either classified as reasonable, questionable, or erroneous in the status_id column or NA if the value is missing.

Usage

classify_time_series_data(
  data,
  ...,
  date_time = "date_time",
  value = "value",
  questionable_min = 0,
  questionable_max = 30,
  erroneous_min = -0.5,
  erroneous_max = 40,
  questionable_rate = 2,
  erroneous_rate = 5,
  questionable_buffer = 1,
  erroneous_buffer = 1,
  gap_range = 5
)

Arguments

data

A data frame.

...

These dots are for future extensions and must be empty.

date_time

A string indicating the column name of the POSIXct vector.

value

A string indicating the column name of the value vector.

questionable_min

A numeric value indicating the lower bound of the questionable range of temperature values.

questionable_max

A numeric value indicating the upper bound of the questionable range of temperature values.

erroneous_min

A numeric value indicating the lower bound of the erroneous range of temperature values.

erroneous_max

A numeric value indicating the upper bound of the erroneous range of temperature values.

questionable_rate

A numeric value indicating the rate of change (temperature per hour) of temperature values that is considered questionable.

erroneous_rate

A numeric value indicating the rate of change (temperature per hour) of temperature values that is considered erroneous.

questionable_buffer

A numeric value indicating the buffer in hours for questionable values.

erroneous_buffer

A numeric value indicating the number of hours buffer for erroneous values.

gap_range

A numeric value indicating the number of hours between two non reasonable values that will be coded as questionable or erroneous.

Value

The original data frame sorted by the date time with a status_id column.

Details

The function only works on a single time series.

The function will error if there are missing or duplicated date time.

The data is processed by:

  1. Classifying the time series values based on their values (`questionable_min, questionable_max, erroneous_min, erroneous_max).

  2. The rate of change to each value is then calculated and the values are classified based on the absolute rate of change (questionable_rate, erroneous_rate).

  3. Adjacent values to all questionable/erroneous are then coded as questionable/erroneous.

  4. Next any value within the time buffer of a questionable/erroneous value is classified as questionable/erroneous (questionable_buffer, erroneous_buffer).

  5. In addition, ignoring the buffer, reasonable values between two questionable/erroneous values are coded as questionable if the hourly duration of the gap is within the (gap_range).

Examples

data <- data.frame(
  date_time =
    as.POSIXct(c(
      "2021-05-07 08:00:00", "2021-05-07 09:00:00",
      "2021-05-07 10:00:00", "2021-05-07 11:00:00", "2021-05-07 12:00:00",
      "2021-05-07 13:00:00"
    )),
  water_temperature = c(4.124, 4.078, 4.102, 4.189, 4.243, 6.578)
)

classify_time_series_data(data, value = "water_temperature")