This repo provides a solution to finding a number of 'key stats' within data in Golang.
To start using KeyStat, install Go and run go get:
$ go get -u github.com/watson-sam/key_statThis will retrieve the library.
For each function offered I will try and flesh out a bit both what I am trying to achieve with the code as well as give an example use case; these are entirely in as sporting context as that is the original point of this library although they should be fairly universal in terms of what they achieve and hopefully the sports context will help make the aims easier to understand.
Definition
For each supplied points how many have been consecutively 1; supplied with either a 1 or 0 'streaks' can be found, for each additional 1 supplied the streak value increases by 1 and is reset to 0 when the supplied value is 0.
If the data supplied is the outcome of a football game with a 1 representing if a team has won and 0 a team has not won, with a streak value of 5 one could say "Team X has won 5 on the bounce".
Usage
package main
import "github.com/watson-sam/key_stat"
func main() {
var samples = [10]float64{
1, 1, 1, 1, 0, 0, 0, 1, 1, 1,
}
so := NewStreakObject()
for _, f := range samples {
so.Add(f)
}
so.Value() // => 3.0
}Definition
The historic method is used to capture the extremities of the data supplied, specifically the historic maximum or minimum values given, this method also includes the functionality to compare a given value with the current min or max with an outcome of 0 meaning supplied value is not equal or more extreme (large or smaller given the what stat was requested when setting up), 1 means supplied value is equal to the most extreme and 2 means that supplied value is the most extreme.
If the data supplied are the historic scores of a football team, if the current stored value is 5 and today they scored 6, when querying the Compare method we would have 2 returned and one could say "Team X has scored 6 goals, this is the most in their entire history".
Usage
package main
import "github.com/watson-sam/key_stat"
func main() {
var samples = [10]float64{
10, 10, 6, 8, 7, 7, 7, 10, 6, 20,
}
ho := NewHistoricObject("min")
for _, f := range samples {
ho.Add(f)
}
ho.Compare(5) // => 2
}Definition A key stat is aiming to find 'meaningful' stats about supplied data; the method looks over all points in the supplied window (starting with the maximum set with maxWindow and decreasing to a limit of minWindow if no meaning is found), what a meaningful stat is given with the cutoff which is the minimum percentage of points in the window that need to be 1. I appreciate this is frankly a terrible definition so hopefully the clouds will clear and the path to meaning will be found with the example.
An example of a key stat could be "Team X have won 5 out of their last 7 games", with the data supplied being binary win (a supplied value of 1) or not (a supplied value of 0) values. With a cutoff of 0.7 and a min and max windows of 5 and 10 respectively.
Usage
package main
import "github.com/watson-sam/key_stat"
func main() {
var samples = [10]float64{
1, 1, 1, 1, 0, 0, 0, 1, 1, 1,
}
kso := NewKeyStatObject(5, 10, 0.6)
for _, f := range samples {
kso.Add(f)
}
kso.KeyStat() // => true, 7, 10
}Currently developing a release process so this will be fleshed out with greater detail in the future, currently we are accepting pull requests for minor fixes, etc:
- Small bug fixes
- Typos
- Documentation or comments
Feel free to open issues to discuss new features.
This repository is Copyright (c) 2021 Sam Watson. All rights reserved. It is licensed under the MIT license. Please see the LICENSE file for applicable license terms.