travishorn / dataset-stats

Dataset Stats

Functions for performing statistics on datasets.

The primary purpose is to run batch predictions. So if you had data like this:

branch	period	sales
A	1	2
A	2	4
A	3	6
A	4	8
A	5	10
A	6	12
B	2	3
B	4	5
B	6	7
B	8	12
B	9	14
B	10	16

And you wanted to predict the sales of new periods, this library can output predictions using linear regression that look like this:

branch	period	sales
A	7	14
A	8	16
B	7	9
B	8	11

Installation

Clone this repository

git clone https://github.com/travishorn/dataset-stats

Change into the directory

cd dataset-stats

Install dependencies

npm install

Usage

This library works best when you already have your data in JSON format, like this:

const inputData = [
  { branch: "A", period: 1, sales: 2 },
  { branch: "A", period: 2, sales: 4 },
  { branch: "A", period: 3, sales: 6 },
  { branch: "A", period: 4, sales: 8 },
  { branch: "A", period: 5, sales: 10 },
  { branch: "A", period: 6, sales: 12 },
  { branch: "B", period: 2, sales: 3 },
  { branch: "B", period: 4, sales: 5 },
  { branch: "B", period: 6, sales: 7 },
  { branch: "B", period: 8, sales: 12 },
  { branch: "B", period: 9, sales: 14 },
  { branch: "B", period: 10, sales: 16 }
];

If your data is in CSV format instead, try using a parsing library like Papa Parse to transform it into JSON. Here is an example:

import { readFile } from "node:fs/promises";
import Papa from "papaparse";

const csvData = await readFile("./input_data.csv", "utf-8");

const inputData = Papa.parse(csvData, { header: true, dynamicTyping: true }).data;
// inputData now looks like the example above

Now import the batchPredict() function.

import { batchPredict } from "./index.js";

Set a list of keys to group on. In this example, we are only grouping on one key: the branch property.

const groupingKeys = ["branch"];

Set which property describes the x values. This is the value that we want to feed into our predictor, in order to have it output a predicted y value.

const xKey = "period";

Set which property describes the y values. This is the value that we want to predict, given a new x value.

const yKey = "sales";

Set a list of new x values for which we want to predict y.

const newXs = [7, 8];

Finally, calculate the predictions.

const predictions = batchPredict(inputData, groupingKeys, xKey, yKey, newXs);

The predictions variable now contains the predicted values:

[
  { department: "A", period: 7, sales: 14 },
  { department: "A", period: 8, sales: 16 },
  { department: "B", period: 7, sales: 9 },
  { department: "B", period: 8, sales: 11 }
]

Notice there is one object per group (each department in this case) and new x (each new period in this case).

If you want your new data in CSV format, you can use a parsing library again.

import { readFile, writeFile } from "node:fs/promises";
const csv = Papa.unparse(predictions);
await writeFile("predictions.csv", csv);

API Documentation

This repository contains full API documentation.

License

The MIT License

Copyright 2023 Travis Horn

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the “Software”), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED “AS IS”, WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.