Before Jupyter Notebook there was IPython. Well it’s actually the precursor and the concepts are one and the same but this is a great quiz for the newbs that are getting introduced to Data Science. But, the real question we’re postulating in this article is if Jupyter Notebook is the new Excel.
I argue that Microsoft Excel seem like it will almost never disappear from financial analysis. It was there towards the modern era of financial analysis and computing and it will likely be around for much longer. However, the question has been raised in the financial analysis and general analyst community do they even need to learn Python or have a Jupyter notebook instance available to them to handle part of their workload. I’d like to talk about three different angles on what we see in the real-world working with FP&A teams, customer engagement analyst, investment banking, and other groups poised to use data to the extreme.
Excel has lot’s of Add-Ins (Advanced Analytics, Solver, etc.)
When you see someone work in Excel that really knows what they are doing, the don’t even use the mouse more often than they have to. It is really a privilege to watch a financial analyst do their thing, crunching numbers in an Excel spreadsheet database with models and formulas that are mind blowing. When it gets data science-like the more advanced users know how to use the MS Excel Analysis Toolpak (https://support.office.com/en-us/article/load-the-analysis-toolpak-in-excel-6a63e598-cd6d-42e3-9317-6b40ba1a66b4) and really create some amazing algorithmic value on their data. Or they’ve had a chance to work with Solver (https://www.solver.com/) and make use of those incredible features. All of which require no python notebooks or really any coding skills to get some amazing compute and analytical power where computer vision, object detection, and to some degree SQL are not necessary.
SAS has been around for a long time too!
Getting involved in Data Science today or graduating university with a data science slant, it is possible for one to not even be familiar with SAS. I wouldn’t be surprised if the youngest generation existing university with a DataScience degree were completely unfamiliar with how profound SAS or MatLab have been prior to the amazing recent breakthroughs in Python, Jupyter, and the libraries upon libraries which are empowering most of the modern data sciences. But those systems are in-fact near ubiquitous and very much still in use. So, even as an alternative, though we see most SAS users or consumers just getting SAS exports to Excel at the end of the day, we see this is a system that already has a deeply skilled workforce working with Excel as an output and not necessarily needing any logic from python or Jupyter notebooks.
Python libraries are the real quest for power
But the real deal is not necessarily about python itself as a programming language. Python is a programming language that appeared almost 30 years ago. That’s a long time. And it has slowly risen to fame with its ease of coding approach and brevity of compilation that gives it a lower barrier to entry than say Java or C++. That being said it is not just the programming language at is core but the myriad of Python libraries for Data Science that help deliver an attainable yet ladder rung obstacle of reaching python data science superiority.
The libraries in python for data science that can be used stand alone or combined can allow amazing discoveries and dissections of data to take place. This is both the beauty and the beast of the discipline which depending on the initiative or the data science task one is charges with could mean a mediocre solution or in the hands of the right expert and amazing insight that truly adds to the businesses value. They question will often be in the end, if the effort was unguided, “could this have been done in Excel?”