Pitfalls and pro-tips for effective and transparent Business Intelligence tools and services data science applications
0 likes33 views
The document offers practical tips and highlights pitfalls for implementing data governance in self-service business intelligence (BI), emphasizing the roles of various stakeholders such as administrative advocates, business owners, and data curators. It discusses the importance of ongoing maintenance for data stewardship, the significance of clear communication, and the need for a structured workflow to ensure effective data management. Key recommendations include building a centralized repository for data resources, conducting regular maintenance, and focusing on achievable solutions while avoiding common pitfalls.
Pitfalls and pro-tips for effective and transparent Business Intelligence tools and services data science applications
- 1. Data Governance For Self-service BI Pro-tips plus pitfalls for real world data stewards Jason Medina Global Decision
- 2. About Me • Name: Jason Medina (UCI Math) • Data roles since 1998 for various size and businesses • Insights from front lines and ground level • Applicable concepts and actions to enable self-service
- 3. Data Governance For Self-service BI Agenda Jason Medina Global Decision
- 4. Agenda 1.Players and Roles 2.Painting the Golden Gate Bridge 3.Figures don’t lie, liars figure 4.Building business rapport to KTLO 5.A genral solution for specific problems •Select •Arrange •Filter •Mutate •Summarize •Exploratory data analysis •Describe missing context •Expose data linkage •Machine inputs •Event or transactional •Human made •Descriptive •Aggregate •Predictive Model Obtain ScrubExplore
- 5. Data Governance For Self-service BI Players and Roles Jason Medina Global Decision
- 6. The stars of our show Adminstrative advocates Executive sponsor – champion data strategy and tactics for internal and external customer success Business owner – oversee business requirements, processes, policy and procedures that leverage data assets for decision makers Technology resources Data technician – executes technical design, development and deployment to productions data solution Data curator – responsible for tactical administration, governance and maintenance of BI infrastructure, resources and user enablement Data consumer – community of users who provide real time feedback in both production and testing environmentsDependency: xkcd.com/2347/
- 7. Data Governance For Self-service BI Painting the Golden Gate Bridge Jason Medina Global Decision
- 8. Masters of our data universe Business deliverable Self-Service BI Tools User curated and supported data sources Enterprise reporting tools IT certified data sources with KTLO production support Spreadsheets Manual inputs as well as flat file exports and other connections “…to ensure the uniformity, accuracy, stewardship, governance, semantic consistency and accountability of an enterprise’s official shared master data assets.” Magic Quadrant for Master Data Management Solutions Published 13 January 2020 - ID G00382085 Ask Want
- 9. Routine maintenance for flexible persistence “There are a couple of misconceptions about how often the Bridge is painted. Some say once every seven years, others say from end to end each year. ..truth is that the Bridge is painted continuously. Painting the Bridge is an ongoing task and a primary maintenance job.” https://www.goldengate.org/bridge/bridge-maintenance/painting-the-bridge/ Things to know: • Be aware of business priorities • Know the reporting calendar frequency and intervals • (Tech)nology is a four letter word Things to do: • Build a “one stop” repository to share maintenance know how • Schedule regular server patching and code reviews • Define problems without words Things to minimize: • Downtime during critical reporting periods • Perfection at all costs • Boiling the ocean “Persistance gets you there, consistency keeps you there.”
- 10. Data Governance For Self-service BI Figures don’t lie, liars figure Jason Medina Global Decision
- 11. Yes but no and it depends (The Rule of 3) The first time they agree to something or give you “yes”, that is #1. Next use a label along the lines of “It sounds like what you want/what you agreed to is X. Their answer to that label is #2. Last, paraphrase what they said – “Please forgive me, I want to make sure I have this right…etc.” – the counterpart’s response to that is #3. You have just executed The Rule of 3. blog.blackswanltd.com/the-edge/what-makes-you-think-your-yes-is-real Things to ask: • What about this does not work for us? • How would this operate in production? • What business objective is achieved? Things to do: • Netrualize negativity by focusing on positive outcomes • Clear path to focus on a sufficient solution • Minimize time to failure Things to avoid: • Ambiguity from pro-noun confusion • More than one optimization goal • Unmeasureable or immaterial outcomes
- 12. Keeping the lights on (KTLO) Things to ask: • Is this a data or an infrastructure problem? • Is this an urgent decision or question? • Has anyone looked at the data? Things to do: • Inventory data with context for business relatability • Clear sky monitoring for stormy Monday alerts • Get to the source Things to avoid: • Data surprises or service gotchas • More than one optimization goal • Lonely watchdogs artcenter.edu/connect/events/toyota-dialogues-fall-2016-edward-tufte.html Edward Tufte speaking at Art Center in Pasadena for Thinking Eye exhibt opening (picture by me Tufte on stage) “Our only language is vision” – Edward Tufted Thinking eye event
- 13. Data Governance For Self-service BI A general solution for specific problems Jason Medina Global Decision
- 14. Workflow automation lifecycle for self service BI Plan Organize Interaction Execute Measure Reality Vision
- 15. Administrative and Technical teams want answers and insights Answers Recommendation actions Detailed explanations and technical documents Business context Technical details Insights TechAdmin
- 16. Data Governance For Self-service BI Thank you Connect at Linkedin.com/in/jason-medina Jason Medina Global Decision
Editor's Notes
- #2: Initial title feels welcoming for beginners Other subtitles: Pitfalls with pro-tips for data policy makers Hi let’s allow a couple of minutes for people to settle in. Today’s presention is live and intended to be interactive. I encourage Q&A thru WHOVA as chat has been disabled in Zoom. The mute all function is NOT used Hello thank you for connecting with us here on the last day of Datacon LA 2020. I hope these past few days have been both exciting and energizing for you all.
- #3: As a junior college student preparing for a transfer into UCI while studying math I gained my first “corporate job” as a point of sale technician. Which means I traveled up and down the 405, 805, and 5 to repair broken cash registers which were quickly evolving into back office data silos full of purchase orders, invoices, timecards, hr and billing details.
- #4: Initial title feels welcoming for beginners Other subtitles: Pitfalls with pro-tips for data policy makers Hi let’s allow a couple of minutes for people to settle in. Today’s presention is live and intended to be interactive. I encourage Q&A thru WHOVA as chat has been disabled in Zoom. The mute all function is NOT used Hello thank you for connecting with us here on the last day of Datacon LA 2020. I hope these past few days have been both exciting and energizing for you all.
- #5: --Scrub left blank because I can’t even begin to imagine all the untidy and tidy data that is ungathered filtered mutated and summarized Here is what I intend to open for discussion here. Players and roles are the persons and teams necessary to get started with self service BI. In many ways when defining what self service BI is and is not I feel like I am trying to bite my own teeth. Self service BI requires a self formation and accountability with resilient persistance to deliver repeatable insightfulness that stabilize as the business adapts.
- #6: Let’s start by talking about the people or types of people needed to structure data governance for self service BI. These players and roles are common in data compaines of shapes and sizes. Very early in my work life I benefited from working across departments which exposed the importance of working together to define and solve business problems.
- #7: Administrative adovacates can be technological resources and vice versa. In most traditional companies, organizations bifuracte first into Operations and IT. Operations includes traditional revenue supporting functions. IT includes more shared services and enables Operations to generate revnue to develop the business. Over the past 10 years, the lines between the two groups are becoming less clear as data literacy and enablement are entrenched in day to day operations.
- #8: With the program set and players identified the next section introduces the idea of joining IT objectives with business goals to drive measureable business outcomes. Much of what happens in self service BI is so exciting because of how data contributes to quantifying situations or prescriptive actions.
- #9: When users asks for a report what they really want is an answer. One of the most common questions is was this a one time event, the start of new trend or some other type of unknown distortion in the data. The quote in the bottom right was taken from the Gartner 2020 Magic Quadrant for Master Data Management solutions.
- #10: Set it and forget is a nice selling feature for BI tools. In application workflows and schedules require human intervention for code reviews and workflow orchestration. When tranisitioning from IT to finance/accounting role the importnance of month end became very clear for a delay in closing the books meant enterprise reporting delays. Delays in reporting are lose lose situations because any news, good or bad, is greeted with uncertainty.
- #11: The book How to lie with Staistics introduced me to this phrase, Figures don’t lie liars figure. Until then I favored “only bad logic can defeat good logic”. The spirit of each is the same. A simple window shift, log scale or regrouping can alter the story presented but not the underlying events plotting the picture. Data art sounds less rigours than data science or self service business intelligence yet applying business rules and logic can feel arbitary to technical minds.
- #14: Initial title feels welcoming for beginners Other subtitles: Pitfalls with pro-tips for data policy makers Hi let’s allow a couple of minutes for people to settle in. Today’s presention is live and intended to be interactive. I encourage Q&A thru WHOVA as chat has been disabled in Zoom. The mute all function is NOT used Hello thank you for connecting with us here on the last day of Datacon LA 2020. I hope these past few days have been both exciting and energizing for you all.
- #17: Initial title feels welcoming for beginners Other subtitles: Pitfalls with pro-tips for data policy makers Hi let’s allow a couple of minutes for people to settle in. Today’s presention is live and intended to be interactive. I encourage Q&A thru WHOVA as chat has been disabled in Zoom. The mute all function is NOT used Hello thank you for connecting with us here on the last day of Datacon LA 2020. I hope these past few days have been both exciting and energizing for you all.