ThinkTWENTY20

  • HOME
  • ERIC'S BLOG
  • FORUM
  • ABOUT US
  • REGISTER
      • ThoughtLeader Gateway
  • Hey! What's New?

RPA – But Eric, What About Python?

on 03 October 2020

Back in May, I had a Blog entry on Python, and whether financial professionals should know how to code in Python. See https://thinktwenty20.com/index.php/blog/478-should-financial-professionals-know-how-to-program-or-data-analytics-accountants-should-understand-them-and-have-toolsets-to-perform-them for a refresher. I mentioned the use of Jupyter Notebook as a means of storing and conveying the code and documentation and facilitating writing, testing and executing the code.

So what does Python have to do with RPA? 

With RPA, I have tools to automate tasks. Python, too.

With RPA, I have the means to get information from Excel, PDFs and other external sources and work with the data automatically. Python, too.

The use of UiPath Studio for “smaller” purposes is free. The use of Python is essentially free.

Python has tons of “libraries” that add additional functionality to work with just about anything. RPA tools are gaining more and more libraries, and often permit the invoking of code just-in-case.

So, what can RPA solutions do that Python can’t? 

  • Provide a limited/no-code environment in which to operate, particularly focused on screen scraping.
  • Integrate with larger suites for collecting business requirements, moving to the field and managing
  • Minimize programming errors with components tuned for purpose

What can Python do that RPA can’t?  

  • There are few limitations to Python; there’s even an RPA library for Python (imaginatively named RPA for Python or RPA-Python) which simplifies website automation, OCR, and interaction with people on the keyboard and mouse. But, 
  • Python is a general purpose tool, so you aren’t starting off with a user-friendly user interface to get the work done.

With UiPath being available for free for smaller uses, it is up to you whether to begin implementing automation within your organization with Python or with UiPath.

Leave a comment

    Comments

    • No comments found
    Powered by Komento
    Forum Recent Posts
    Data For Sustainability Disclosures
    Monday, 11 September 2023

    By Gerald Trites
    Industry generally, because of legislation and stakeholder demands trying to move towards net-zero carbon emissions. Measuring progress in this endeavour calls for data that may not be available or may not be consumable because of formatting and general system disparities. Companies are working to rectify this situation and gain ready available to the needed data. It can be a big job.

    Equipment being used for production, travel, deliveries etc. that generates carbon emissions often is older and does not have the sensors needed to measure emissions. It may not be practical to replace that equipment so add-ons are often needed to capture it. That will take time to resolve.

    In the meantime, companies are left with finding ways to do useful estimates. The data – actual and estimates - needs to be fed into their processes and systems to be useful.

    Companies are working on this, especially utilities and energy companies, mostly on scope one emissions. Scope two and three emissions are even harder to deal with from a data viewpoint, because they are outside the direct control of the company.

    Check out this article which explains the issue well.

    The Need for Better Awareness of AI's Downfalls
    Monday, 21 August 2023

    By Gerald Trites

    “A recent study by the Canadian Institute for Advanced Research (CIFAR) shows that Canadians are diving into new AI tools with excitement but are missing out on understanding the potential downsides.”

    Especially since the advent of ChatGPT, the use of AI in various arenas has been expanding quickly. Everyone is talking about it but not everyone is aware of the issues it presents. This is slowly changing, since many people are experimenting with Generative AI and learning that it doesn’t always produce reliable results. In fact, sometimes it invents its own inaccurate results. This can be annoying but becomes a serious problem if the AI is being used, for example, for medical diagnosis.

    ThinkTWENTY20 © 2023. Privacy Policy