Why Do Machine Learning in Excel?

If you want to wade into the world of machine learning, there are open-source solutions that will usually meet your needs.  The question arises as to why anyone would do machine learning entirely in Excel?  If you were to go to Reddit or Stack Exchange and ask a question about it, you should be prepared to be dissuaded from even suggesting the idea.

There are two things I know for sure.

  1. A lot of people use Excel for a lot of things.  Many of them probably not what the inventor of the spreadsheet ever intended or imagined.
  2. A lot of people are too busy with their real jobs to code or learn one of the major machine learning platforms.  Since they aren't using it every day, going back to a second machine learning project is going to involve relearning how to use a platform all over again.

These two notions were the genesis of Vaimal, my machine learning add-in for Excel.  Machine learning is only going to get more pervasive in business and life in general.  It was my goal to help bring machine learning to a more general audience such as small businesses and family farms.

What Vaimal Doesn't Do Well

It may seem odd for me to discuss weaknesses in my product, but I'm going to address them anyway.  Vaimal is not for big data and it will take longer to train machine learning models than platform X.  Vaimal is not intended for data scientists.  It's for the occasional user where time for model training isn't the first priority.

Where Vaimal Excels (no pun intended)

  • Everything is in Excel.  You work in a familiar environment.
  • Again, everything is in Excel.  Once your data is brought into Excel it never leaves.  There's no sending the data to the cloud for processing.  If you can import the data into Excel securely, then it's safely in one place.
  • There's no coding required.  All operations are performed through a user interface.

What's Required of You

Vaimal contains five basic machine learning models and two ensemble methods.  With this many choices, you need to decide which one is best for your problem.  Machine learning requires some experience and intuition to tune models, but there is also a lot of trial and error involved.  This requires at least a basic understanding of how each model or ensemble learns and the factors that affect learning.  The good news is that you will be learning about the models themselves instead of worrying about coding and correct implementation.

Don't worry though, the Vaimal user manual contains background information on each model and ensemble to get you started on your machine learning journey.  This website also contains many articles on machine learning topics to help you along.  Check out the Knowledge Base for more information.

Give Machine Learning in Excel a Try

I encourage you to download the free trial version and start playing around with it to see if it's right for you.  You can get your copy from the Vaimal product page.  I also encourage you to subscribe to the Vortarus newsletter to get the latest info on machine learning topics.