Module 1

 Module 1!

Hi everyone!

We’ve made it through another few weeks. In this blog post I have made remarks and analysis on every video lecture since the last module. I paid special attention to our latest lecture on data dashboards. I also included a link of my own research for every video lecture. I hope you all have a chance to read one of these articles and give me your own opinion and arguments. Let me know if you agree with the author.

 

 

Data Warehouse:

Management decision making is some of the most critical and vital aspects of any company. Having easy access to high quality information about the process is vital to making those decisions. This lecture is mostly just terminology. I agree with the importance of creating a data warehouse and the process seems to be based around using existing well-designed databases. I can say however, at my work we use reporting tools & analytics all the time to make important discussions. We base the size of our servers based on usage reports and we also focus on improving elements that see a high amount of traffic. We have also started doing datamining on some minor projects. I hope that we can get more into datamining in the form of machine learning to assist us in presenting data. Hopefully after I finish this degree, I will be free to go study the mathematical concepts behind that sort of work.

 

In my research on this subject I found an article on the key elements of a modern data warehouse. This article written by David Weldon compares modern data warehouses to the traditional older data warehouses. In the article Weldon argues for modernizing the old data warehouse. One of his most compelling arguments is that a modern data warehouse allows for real-time updates. I could see this being a vital part for the decision-making process.

 

https://techbeacon.com/enterprise-it/key-elements-modern-data-warehouse

 

 

 

Balanced Scorecard:

Admittedly this was one of the lectures I really struggled with. The quiz for this lecture was particularly hard for me as many of the concepts seemed to overlap and I was having a hard time differentiating them. The components of the score card all make sense to me, but they all seem to overlap at some level. I wonder if there is a version of the score card but with overlapping regions of responsibility. One of the questions on the quiz that got me was specifically asking about financials so I just assumed it would be a financial sector. I guessed wrong. The idea of strategy is something that I understand the importance of. In some of my previous classes I’ve seen many patterns of businesses that failed to stick to a strategy and ended up going out of business. When designing any sort of IT infrastructure or really any infrastructure in general it’s always important to align yourself with the strategy of the company. Southwest is a great example of this, and I’ve noticed it’s listed in many lectures/videos/articles when it comes to strategy. I remember when I first heard about southwest airlines, I was immediately impressed by their business strategy.

Harvard Business School recently released an article written by Robert Kaplan et al that analyses the changes needed for balanced scorecards in the modern market. Some of the changes they recommend are replacing the financial sector and replacing it with ‘outcomes’. This is to encompass additional metrics including societal changes. Additionally, the customer sector is replaced by ‘Stakeholders’. This is to account for the multiple stakeholders that may be invested in the company. A central theme among the report. I would recommend reading some of the conclusions of the article as it could be vital in the modern company.

 

https://www.hbs.edu/faculty/Publication%20Files/WP21-028_Triple_Bottom_Line_8.28.20abstract_correction_dd38a54c-48f2-4471-80db-e0ed6f962309.pdf

 

Star Schema Design:

 

This lecture was really interesting, and I learned a lot about advances techniques in star design. I particularly found interesting the surrogate key topic. I suppose I loosely understood the concept, but I had never conceptualized the importance in my actual work. I feel like certain elements of database design that exists around database transitions are hard to prioritize until you’ve actually been in that situation. This lecture covered a number of other dimensions that make sense in the scheme of database and/or star schema design. Slowly changing dimensions seem particularly important. Making the decision on how to transfer data is something that can make or break business intelligence. Hopefully a good database design would prevent this problem from ever happening, however. In my database class we learned about different forms for databases. Some of the later normalized forms would prevent some of the problems with transitions present in this lecture.

 

I know that database design isn’t exactly star schema design, however they do share a ton of similarities. I found a recent article about tips & tricks with database design and one of the top tips really resonated with me. Rice writes “don’t be a perfectionist and do think in tradeoffs.” I think this is important as you’ll never design a perfect database. It’s simply impossible.

 

https://builtin.com/data-science/design-a-database

 

 

Data Quality Analysis:

 

I agree wholeheartedly that data quality is super important to the business process. I deal with data quite a bit. Recently we had a problem where we allowed people to enter their own addresses into a web form. Huge mistake. 90 percent of individuals were able to enter a valid address. However, for the remaining 10 percent we got a wild array of “addresses”. Things like email addresses to incomplete addresses were a plenty. We eventually solved this by running all addresses across the USPS before inserting them in our database. However, in the meantime we lost a ton of time trying to normalize and validate addresses. We did do some profiling to clean the bad addresses, but it wasn’t as organized as the steps listed in the lecture.

 

In researching data quality news, I found this article that talks about the idea of standardizing and grading the quality of big data. The lack of standardization in big data has created an entire industry of data cleaners. Imagine if you went to buy a data-set and it was graded similarly to how our meat is graded. The dataset is full of nulls? That’s a some low quality data. I recommend reading the article, at least the part of the article where Prokop lists the dimensions of grading. One of her first suggestions is “Accuracy: Is the data correct?”

 

https://streetfightmag.com/2020/09/15/standardizing-the-definition-of-data-quality/

 

Dashboard Design & Analysis:

This was the majority of my time this module. Learning Tableau took forever, and while a lot of the concepts in tableau were self-explanatory, many were not. I took a data visualization class in my undergrad and I’m thankful that I did as I feel like many concepts were able to be cross applied here. As a summary, this module was all about making an effective dashboard. Further, it focuses on ensuring that every inch of your dashboard is useful and follows certain guidelines. Additionally, this lecture talks about the different types of dashboard. My project fell mostly under the analytical dashboard. I think the overall concepts of dashboards are super interesting. At my current job we make a ton of visualizations and admin panels. However, I don’t think I’ve ever had a situation when we were requested to combine them. I can certainly see the appeal for certain situations, however. I watched the TED talk associated with this video and the main message I got from it, is that data can be misleading. Well that, and the fact that we have pretty naïve understandings of populations and cultures around the world. However, Hans Rosling was able to demonstrate both good and bad ways to display data which I thought was remarkable.

 

From the reading from this week, most of it to me was just an extension of the lecture. However, I did notice something pretty funny that I wanted to share. Stephen Few made a brash prediction back in 2005 that data dashboards wouldn’t be web based in 10 years. This obviously isn’t true as the majority of data dashboards are now web based. Web technology is amazing and it’s truly getting better and better every year. I’m certainly not a fan of how bloated everything is getting, however, I think this is just another problem that will eventually be solved. Creating a data dashboard seems the easiest in web technologies as it’s easily transferable to any device. Even our Tableau dashboards are web technology based.

 

For my external research I was attempting to find some recent news articles or analysis on data dashboard. From my research it actually seemed like data dashboards are falling out of popularity. However, the rise of Covid-19 has brought back the data dashboard in full swing. A quick search will show you that a number of universities and cities have created their own data dashboards for tracking the virus. My guess is that important board members & politicians need to be able to make important decisions regarding the virus. For them to do that, they need all the data in one place. One such example is from Texas where a university built a dashboard for the entire state. The article remarks on how important the dashboard is to the decision making process.

 

 

https://edscoop.com/university-texas-launched-statewide-covid19-dashboard/

 

 

Citations Proper:

 

Foresman, Betsy. “University of Texas Built State's COVID-19 Dashboard.” EdScoop, 25 Sept. 2020, edscoop.com/university-texas-launched-statewide-covid19-dashboard/.

Kaplan, Robert S. “Updating the Balanced Scorecard for Triple Bottom Line Strategies.” Harvard Business School, 2020, www.hbs.edu/faculty/Publication%20Files/WP21-028_Triple_Bottom_Line_8.28.20abstract_correction_dd38a54c-48f2-4471-80db-e0ed6f962309.pdf.

Prokop, Kristina. “Standardizing the Definition of Data Quality.” Street Fight, StreetFightMag, 15 Sept. 2020, streetfightmag.com/2020/09/15/standardizing-the-definition-of-data-quality/.

Rice, Mae. “Designing a Database: What You Need to Know.” Built In, 10 May 2020, builtin.com/data-science/design-a-database.

Weldon, David. “The Key Elements of a Modern Data Warehouse.” TechBeacon, TechBeacon, 8 Sept. 2020, techbeacon.com/enterprise-it/key-elements-modern-data-warehouse.