TimeSeries is a typescript object designed to keep data and perform calculations on date and value timeseries.
The data is stored in an array of TimeSeriesItems, where TimeSeriesItem is basically
{
timestamp: Date;
value: number;
}By default the TimeSeries class imports json data in the following format.
Example
{
"key": "any id or name (optional)",
"timestamps": ["2016-08-31", "2016-09-30", "2016-10-31", "2016-11-30"],
"values": [1.23, 1.27, 1.24, 1.31]
}Where:
- key: String, any id or name, optional
- timestamps: Array of date strings
- values: Array of numbers
- clone(): Creates a deep copy of the timeseries
- assign(): Replaces all properties in the timeseries
- range(startDate, endDate): Like Array.slice create a sliced copy of the timeseries, from startDate to endDate
- sort(): Sorts TimeSeriesItems in ascending date order. Which many functions and operations require. Therefore items should always be kept in ascending date order.
- values: Array of numbers
- timestamps: Array of Date
- start: First timestamps
- end: Last timestamps
- startValue: First value
- endValue: Last value
- count: Number of items
- periodicity: Number of items per year rounded to one of the standard values 252 (banking daily), 52 (weekly), 12 (monthly), 4 (quarterly), 2 (semi annually), 1 (yearly)
- indexOf(): Returns the last index in items, where timestamp >= item.timestamp
- latestValue(): Returns value of item found by indexOf()
These functions / operators modifies each item value and returns a ref to the TimeSeries object itself (for method chaining)
- log(): Log of each value (safe log, returns Number.NaN if non positive)
- exp(): Exp of each value
- add(v): Add v to each value
- mult(v): Mult v to each value
- neg(): Negates each value
- inverse(): 1/x of each value
These functions returns a timeseries one item shorter and do not modify the object itself
- diff(): Returns a timeseries one item shorter and with item to item differences
- return(): Returns a timeseries one item shorter and with item to item quotients (returns)
- logReturn(): Returns a timeseries one item shorter and with item to item log quotients (log returns)
These functions returns a timeseries and do not modify the object itself
- cumSum(): Returns a timeseries with cumulative sum values
- cumProd(): Returns a timeseries with cumulative product values