NStack provides an library of data science functions which you can run on your imported datasets. A function is a piece of code which allows you to do everything from data exploration and wrangling to predictive modelling and machine learning.
Functions are written in Python or SQL, and are typically open source in order to provide transparency and allow discussion and improvement, but you do not need to know how to to code or have a background in software engineering or statistics to use them on your datasets. They are contributed by either the NStack team, the NStack community, or can be published and shared internally inside your company.
You can see a list of NStack's available functions at https://app.nstack.com/functions/, which is split into two categories:
- Public Functions are functions which are contributed by NStack's community of data scientists. If you publish a function on NStack and set it to Public it will appear in this list for other NStack users.
- My Functions are functions -- either public or private -- that you have published under your own account. To publish a new function under your account, hit the Publish New Function button. This will only be available if you have logged into your NStack account.
Viewing a Function
Once you browse to an individual function, you will see the following sections:
- Schema: this outlines the input and output schemas of the function; that is, format and type of data that the function can take as input, and the format and type of data that it produces as output. If the input schema of your dataset does not match the input schema of the function, you will get an error and your function will not be run, so you should use the Dataset Transform functionalities to get your data in the correct shapre. Some functions have an
Any datasetschema which allows all formats of data.
- Description (Optional) An in-depth "Readme" of the function
- Runner: this is where you select your input Dataset, and where output is sent.
- Source Code: the source code of this function
Using a Function
To make use of a function, you can load it into your Notebook, where you can run it on your Datasets. If you are logged in, a function's individual page will have a shortcut to Add to Notebook.
Creating a New Function
You can create a new function from the notebook view. Once you have made a successful run with your function in the Notebook, click the Save Function button. This will take you to a page where you can enter a description, title, and headline for your new function.