This group is about innovative solutions for Microsoft Excel. It will focus on the different use cases of xlwings ( https://www.xlwings.org ), the open-source solution connecting Python and Excel. We are, however, also open to any other solutions and ideas that help getting more out of Excel. We are especially interested in solutions that make Excel more robust and connect it to a modern tool set (like machine learning, continuous integration etc.).
Talk 1) Using in-memory objects to integrate Python in Excel, Julij Jegorov
The excel add-in utilizes xlwings technology to create in-memory nested key-value pair objects.
The object parameters (numpy arrays, pandas dataframes, instances of python classes, etc...) can be referenced directly in Excel spreadsheets and used as input or output arguments
to Python functions or serve as functions themselves.
The use-cases include:
- Scripting Python functions directly in Excel cells.
This allows utilization of Python libraries like numpy, pandas, scipy, etc... without an IDE.
Example: integrate Python numerical functions that are not readily available in Excel.
- Loading Python classes into spreadsheets and running selected functions.
State variables can be passed as class arguments and shared by functions within the class.
Example: when pricing financial instruments passing 'calculation date' as a class variable ensures that all relevant functions are re-calculated when the date changes.
Talk 2) Deploying Machine Learning Models via Excel for Traders, Colan Walsh (Founder at Achilleon Consulting)
For all its flaws, Excel remains the standard tool of wholesale traders in the financial industry. Excel trader workbooks have a great variety of forms: risk management sheets, RV trade screeners, historic market data analysers, etc.
Two commonly observed features of trader workbooks:
1. They incorporate one or more statistical models; from simple linear regressions, to complex Monte-Carlo-based simulations.
2. They are the result of years of effort; re-engineering them from scratch in another language is often not practical, or desired by the end user.
In this talk, I will discuss how to easily upgrade the simple statistical models in trader workbooks to more powerful Machine Learning models (including neural networks) without extensive workbook re-engineering.
Talk 3) Deployment of xlwings powered spreadsheets, Felix Zumstein (Creator of xlwings)
One of the biggest hickup in the Python for Excel story is deployment. So let's have a look at our options and pro's and con's.