This document summarizes a presentation about using Excel and Python together for financial analysis. It discusses: 1) Using Python for interactive data analysis and automated testing of Excel models. 2) Creating Monte Carlo simulations and exporting Excel macros to web apps with Python. 3) Developing user-defined functions in Python to extend Excel's functionality, like creating a correlation heatmap from multiple data sets.

1. Excel & Python
• Open Source Technology in Fintech, Zürich
July 14th, 2016 Felix Zumstein, CFA

2. Material
https://github.com/ZoomerAnalytics/talks/tree/master/SwissFinteCH_zrh_20160714
2

3. About me
• Consultancy (Zürich):
– Consulting: Web apps for Finance
– Workshops: Excel & Python
– Open-source: xlwings
• Previously:
– 9yrs in Banking /Asset Management
– Background: Finance & Economics
3

4. Agenda
4
1) Scripting/Interaction
• Interactive data analysis
• Automated Testing of Spreadsheet Models
2) Macros
• Monte Carlo Simulation
• From Excel to web apps
3) User Defined Functions (UDFs)
• Correlation Analysis

5. 1 Scripting/Interaction

6. GDP per capita
6
Source of xls file:
http://data.worldbank.org/indicator/NY.GDP.PCAP.CD
+

7. Spreadsheet Risk: Automated Testing
7
• Audit/Regulation require effective controls
• Automated unit tests are a cornerstone of modern
software development (”test-driven development”)
• Excel: doesn’t offer any means to create unit tests
• Python: has a built-in unit test module
• Unit testing with xlwings is unintrusive:
- No add-in to be installed
- No changes to the original VBA code base
- Python standard lib + xlwings is all you need
• It’s a great way to verify & debug legacy VBA code

8. 2 Macros

9. Monte Carlo Simulation
9
simulation.py
xlwings_app.py web_app.py

10. Source code and hosted app
10
Hosted Sample: www.zoomeranalytics.com/simulation-demo
Source Code: https://github.com/ZoomerAnalytics/simulation-demo

11. Prototype analytical web apps
11
vs.

12. 3 UDFs

13. DJIA: Correlation Analysis
13
=CORREL(array1, array2)
Excel’s Correlation formula accepts just 2 data sets:
Here’s how we’re going to fix this:
• Add CORREL2 to get the full correlation matrix
• Visual representation: Heatmap

14. Questions?
twitter:
@zoomeranalytics
xlwings newsletter:
via xlwings.org
www:
xlwings.org, zoomeranalytics.com