Tag: Analytics

Python DoWhy

Microsoft’s DoWhy Library for Python greatly simplifies the task of estimating causal effects.  If you or someone you know is involved in data analysis, it is worth your while to see what DoWhy can do for you.

I have spent a little time working with the library and although I have no coding examples at this time, the powerful nature of this library prevents me from waiting to share it.

Visit the DoWhy github page for more details on the DoWhy library.  The information and documentation presented on the site provides sufficient detail to download and start working with the DoWhy library.

Until Next Time – STAY lean!

Versalytics-Logo

Related Articles and Resources

  • https://github.com/Microsoft/dowhy

 

 

Advertisements

Pure Genius: Teach, Build, Model

At first I thought this approach seemed to be a little too simplistic as an execution strategy.  On second thought, however, it seemed to be very appropriate to most situations that we may encounter.  It has certainly been a very practical superview for our current project.

Teach: Teach the concepts, systems, processes, and procedures that are required to achieve a given goal or objective.

Build: Interactive training sessions should provide for feedback and further development of the elements presented.  At this stage, personnel become engaged and help to further develop the systems, processes, and procedures.

Model: Leadership, executive management, and personnel must model the behavior or activities that are required to successfully achieve the desired goal or objective.  Deviations to the planned activities must be discussed and reviewed in real-time with the team.  Revise as necessary but enact or communicate changes in real time.

How far do we go in our pursuit of better systems and data analysis? The true goal is not to build systems that collect data for later analysis, but to build systems that analyze data based on collected data in real-time.  Ideally, the analysis would lead to future projections based on past history and current performance in real-time.  How far we want to go with the analysis is driven by the vision and goals we are seeking to achieve.

For added inspiration as to what can be accomplished with data that is available, we found this video featuring Stephen Wolfram, creator of Mathematica, where he talks about his quest to make all knowledge computational — as able to be searched, processed and manipulated. His new search engine, Wolfram Alpha, has no lesser goal than to model and explain the physics underlying the universe.

This is an amazing video (unfortunately, we can’t embed the video directly into our page) and the WolframAlpha Website is equally impressive to demonstrate what can be done with the right tools and data engines. Click here to visit WolframAlpha. While you’re at the site, you may wish to subscribe to TED. There are many motivating and inspirational videos / talks that cover a diverse range of topics and areas of interest.

We are currently developing a business operating system for a client where informal, undocumented, methods exist.  Our task is to streamline these “methods” into a cohesive system to provide for real-time analysis and reporting for a variety of functional areas in the business.  Essentially, we are creating a bridge from the present day activities to a more rigorous, systems driven, process.  From this perspective, the above three (3) step process has proven to be a very practical means of strategizing the approach as new initiatives are introduced.

Until Next Time – STAY lean!