Me and spreadsheets go way back. My first job was hacking Lotus 1-2-3 at Kodak as a summer intern in 1985. I consider it my first programming job.
A year later, I got interested in neural networks, and built back-propogation models in Excel on the mac. Yes very SLOW but still a great way to learn.
As side project ten years back was to leverage Excel's native web query (IQY) mechanism to build a profitable SaaS company just based upon letting users get data into Excel from various 3rd party social media and analytics platforms.
Now I work in big oil and our team basically turns Excel models that users create into scalable data warehouse apps.
Even after years working with Excel, I still consider myself a journeyman.
Using excel for learning NN. Multi input on a node would be a bit wonky (but doable) and layers could just be sheets in the workbook. It even already has a bunch of nice built in tools to import/export data. That is one of those 'why did I not think of that' things!
A year later, I got interested in neural networks, and built back-propogation models in Excel on the mac. Yes very SLOW but still a great way to learn.
As side project ten years back was to leverage Excel's native web query (IQY) mechanism to build a profitable SaaS company just based upon letting users get data into Excel from various 3rd party social media and analytics platforms.
Now I work in big oil and our team basically turns Excel models that users create into scalable data warehouse apps.
Even after years working with Excel, I still consider myself a journeyman.