Thursday, May 7, 2015

Gartner EIM / MDM Summit 2015 Conference Attendee Summary (with Tweets)


The Gartner Enterprise Information Management & Master Data Management (EIM / MDM) Summit was held April 1 - 2, 2015 in Las Vegas, Nevada.
From the Gartner events website:
The Must-Attend Event for Information & Data Management Professionals. Enterprise information and master data management can help organizations perform more efficiently with cleaner, more trusted data by creating a single view of customers, products and suppliers. But how do you get started? And if you've laid a foundation, how do you become world-class? Designed to address all maturity levels, Gartner Enterprise Information & Master Data Management Summit delivers the insights, frameworks and best practices across information management topics to help you take control of your data and dramatically improve business performance.

Overall Impression

Over the course of two days, I attended 10 sessions focused on MDM and EIM. I tweeted a lot - I sent a total of 219 tweets highlighting key takeaways from each session. (Tweets are below.)
I had a few opportunities to network with fellow attendees, but it was obvious to all of us that we were meant to focus on absorbing the knowledge of the Gartner team rather than comparing recipes with one another.
In all, I derived a lot of benefit from the conference. I learned a lot and felt it was absolutely worth my time to attend as an end user of MDM systems and manager of data governance initiatives. Based on this experience, I am a firm believer that every organization needs to perform a data assessment to understand what they have, what shape it's in, and where the value lies. Once that is complete, an organization can organize, cleanse, manage, and monitor, improving their ability to make strategic decisions.


  • EIM = Enterprise Information Management. All data in your org.
  • MDM = Master Data Management. 
  • Master Data = Data that changes infrequently enough and that is business-critical enough to capture and "master" 
    • Generally, master data == source of truth
    • Generally, master data != transactional data
  • DQ = Data Quality 

Sessions Attended

Gartner Opening Keynote: Maximize Business Value With Your Digital Information Strategy

  1. Presenters:  
    1. Ted Friedman, VP Distinguished Analyst, Gartner  (Twitter handle: @ted_friedman)
    2. Debra Logan, VP & Gartner Fellow, Gartner
    3. Andrew White, VP Distinguished Analyst, Gartner (Twitter handle: @mdmcentral)
  2. Summary:
    1. Information is a key ingredient for any digital business. It is the one thing that is constantly being exchanged between businesses, between people and between things. In effect, information becomes the critical connection that links together the value chain of organizations. As all industries make the transition to the world of digital business, organizations can maximize value from better information management practices through innovation, value creation, efficiency and risk mitigation. An information strategy tuned for the challenges of digital business is the key to success.
  3. My notes (via tweet)
    1. "We will see a shift away from asset ownership to asset access." ~ @mdmcentral
    2. "Since the mainframe era, only 50% of orgs have followed a rigorous evaluation of their IT environments." @mdmcentral 
    3. "We need to move to a state where we are constantly evaluating and showing the value of IT/systems to the business" @mdmcentral
    4. "Digital deficit emerges when we budget for mediocrity. Our task is to combat this!" @mdmcentral 
    5. Questions for the business: What is most important strategic information we have? Depends on the biz outcome we seek!
    6. Its never too late to stop projects that have no "why"!
    7. First mandate: change the way you speak. Have an answer to the question "what outcome do we derive from our information"
    8. Why are we doing this? What will it get us? Which processes will be affected? These define success on a business level.
    9. CIO's responsibility is to ignore traditional IT metrics and focus on optimizing biz information to answer strategic questions
    10. Governance is simply the rules by which things are done, who makes those rules, and how they are enforced.
    11. Bimodal governance = big governance (regulatory) + agile governance.
    12. Regulated information + vital records -> Big governance. Business continuity + emphemera (90% of your data) -> agile governance
    13. Agile governance = simple rules & easy - to - understand implementation
    14. What is the one thing so important that you can't NOT do it? The most successful orgs will be the ones who focus on their people
    15. Everyone who engages w information - internal, customers, vendors - can benefit from open structures with well-defined processes
    16. "Even the office trash is handled better than the information - at least it gets cleaned up every night!" @ted_friedman
    17. Chief Data Officer manages data for the biz. Creates relevant strategy. Executes on the strategy. Ensures biz value is derived
    18. Data / information management is a team sport. @ted_friedman
    19. Technology should NOT be at the center of Information Management strategy. PEOPLE should be. Humanism is key. @ted_friedman
    20. The data humanist ensures participation by all, encourages experimentation, and creates systems to serve biz @ted_friedman
    21. Avoid "ownership" and focus on "stewardship." Use words like "curate," "share," "innovate," "enablement" @ted_friedman
    22. Intel uses key metrics to understand investments, improve the biz through measured review based on strategic goals @ted_friedman

IQScoring: How to Assess and Advance Information Quality in Your Organization

  1. Presenter:  
    1. Robert Hunter, Enterprise Quality Strategist, Intel Corporation
  2.  Summary:
    1. IQScoring is a concise Information Quality (IQ) assessment methodology that can be applied to key company data sets AND applications. Applying the method creates baseline measurements that are effective for understanding and communicating the current state to executives to impact investment decisions. We’ll describe the IQScoring Method, how it works in conjunction with your business processes and industry recognized frameworks. Quality rules!
  3. My notes (tweets): 
    1. At Intel, we use the data: let's see the data, and we'll make the decision.
    2. Key learning: if Corporate Quality Network partners with groups incl IT, can have discuss w cust w/o "tech speak"
    3. Intel CQN created a survey with APQC to understand their methods. Findings were useful/unsurprising, mostly.
    4. Out of results from the findings of survey, created IQScoring Scale. BI has own scoring range, as does Legal. 
    5. Intel found that BI/Legal data requires very high quality in order to implement systems around them.
    6. Intel discovered if data didn't meet quality standards required for different uses, there would be low-to-no chance of success 
    7. IQ Score is generated from a matrix of questions on areas like People, Data Quality, etc. 
    8. Out of the IQ Scoring comes artifacts like: list of all subject areas at company, know who knows what, and how its used
    9. If the survey shows there are 50 applications that use a certain data area easy to make biz case that quality is vital
    10. Once survey is complete, improvements made, re-run survey to get new IQ score and iterate again
    11. The applications that use the data can only be scored as high as or LOWER than the quality of the data it uses. 
    12. Which biz process uses the data? Who on team works on data quality? IQ (information quality) can't improve w/o answers 
    13. Data quality (a.k.a. DQ) is center of biz. Applications can only work as well as their input/output. IQ score brings shines light.
    14. Data quality *then* Governance *then* MDM?
    15. People are critical. Subject expertise, retention, use data for job, training, train others/evangelize? Q's on IQ score
    16. DQ checklist: complete, valid, accurate, precise, duped, equivalent, timely, presentation clarity, auditable, trusted
    17. How is IQ score being used? Educate company, leverage gate-keeper teams, align with other enterprise efforts
    18. IQ Scoring allows them to advise on business process as it depends on & affects data quality
    19. Now that the IQ Score exists across all orgs, can measure & understand that the needle's being moved by the efforts
    20. Enterprise taxonomy <> IQ Scored data areas <> Enterprise-wide process framework <> accountable people <> corporate orgs
    21. Looking at top 3 levels of classification framework to determine where to focus. Try not go too deep to remain focused
    22. If team has chosen to use own data source & not DQ-approved set, IQ score tanks. Starts convo re address customer need
    23. Quality is concerned w governance. DQ team draws connection btwn biz impactful things (legal compliance) & show problem
    24. How do you get general management buy-in? Work w/ them on the project! Educate management & gate-keepers. Then reinforce!
    25. Weave the process of quality into the fabric of the biz. DQ team spots trends, can engage
    26. When is IQ Survey run? Yearly for sources, at significant launch for applications. 

Informatica: Designing Complex Operational and Analytical Multidomain MDM: Quest Diagnostics’ EIM Blueprint

  1. Presenter:  
    1. Jason O'Meara, Directory of Enterprise Information, Quest Diagnostics
  2. Summary:
    1. Can MDM make a difference between life and death? Well, in healthcare, it does. Quest Diagnostics built the most stringent operational and analytical multidomain MDM that exceeds the rigor of healthcare requirements. In this session, learn how to successfully create an EIM blueprint comprising of master data management, data quality, data security, metadata management, data architecture, and governance. Jason will discuss what worked and what didn’t and how to mitigate risk and improve success.
  3. My notes (tweets):
    1. Change in healthcare from treating sick people to managing health
    2. Data security, data quality, metadata management, MDM, data standardization, data architecture, governance
    3. Managing all facets means that IT/infra is critical part of data quality. Interoperability is basic need.
    4. Multi-client domain model in MDM. In #healthcare the customer is a fragmented domain
    5. Domains have multiple actors == patient, providers / physicians, contacts, clients / employers, providers
    6. Hospital w/ multiple labs, outpatient surgical A, outpatient surgical B. Structure means 1 hospital = n accts
    7. From account-centric legacy structure (traditional) to client-centric structure (related entities are linked)
    8. Don't make it too complicated. Some systems can remain in situ (employee master == PeopleSoft)  
    9. Using party role based system via Informatica MDM to express these relationships in a flexible way
    10. Assume 5 - 10% of records to go into queue in MDM Hub. How many people does it take to fix them? #Headcount
    11. Records may come from a doctor who wrote on paper, then OCR'ed. Have to consider source for quality steps
    12. Evaluate data in each domain. Who has most stake in data quality? That group may own governance for that data
    13. Quest Diagnostics MDM journey: Foundational model>mastering rules>master data validation>integration backbone>ent. integration
    14. Having multiple partners in the business echoing the need to cleanse data makes request for headcount easier
    15. Storing some almost-big data in their holistic repository! 1.5b patient encounters in clinical data repo
    16. Clinical data > expansive test menu > provider insights > patient insights > clinical workflow integration
    17. Putting test results for patients into analytics chain allows point-of-care / clinical workflow enhancement

Chief Data Officers: Who They Are, What They Do and Why This Role Will Keep Growing

  1. Presenter:  
    1. Debra Logan, VP & Gartner Fellow, Gartner
  2. Summary:
    Information is a competitive differentiator, yet this vital resource remains unmanaged in most companies. New roles relating to the management of information, such as the chief data officer, are proliferating and new kinds of information leaders are emerging. 
    1. Who are today's CDOs and what kinds of companies have them? 
    2. What do CDOs do and how will they work with the IT organization and other parts of the business? 
    3. Why will the role keep growing?
  3. My notes (tweets):
    1. "CEO's recognize data as a corporate asset." Do they really? No! Info is not treated as an asset by the org
    2. If data is an asset, we should have a vision/strategy/process around it. This needs to be recognized by CEO
    3. Also seeing another role - Chief Analytics Officer. Today, likely similar role.
    4. McKinsey recommended CDAO. Clearly this needs to be sorted out.
    5. Develops data strategy, governance, control, policy development, and effective exploitation. Should be sr. lvl
    6. Ideally, CDO is a C-suite role that has reporting responsibility to board.
    7. CDO not a tactical role. Has to be strategic. Someone above line-of-biz level to optimize, strategize
    8. 25% of CDOs are women.
    9. Biggest benefit to the business is to exploit data company wide to make money.
    10. Role combines accountability & responsibility for info protection/security + info governance, data quality
    11. Data monetization should be attacked, info security & data quality can be delegated.
    12. 36% of banking co.s polled had a CDO. 16% of Govt agencies polled had a CDO. Up & coming: Healthcare & Insurance
    13. HSBC, Seattle Hospital, UK Govt Office, depts of US Federal Govt have CDOs.
    14. How do they work w/ the biz? Existence of CDO may require CIO to adjust role/expectations
    15. CIO generally doesn't have power to control data in the org. They own systems that hold data, but not bits themselves.
    16. The biz controls the data. The CDO is the biz leader who helps org make decisions regarding data management
    17. Data is a largely unstewarded asset in most orgs. That needs to change, from strategic perspective.
    18. CIO & CDO work together to sort out needs for the technology used and the data management strategy.
    19. CEO needs to take stance that data is an enterprise asset. CIO must be pro-tech and embrace data mgmt
    20. Drives vision, strategy, metrics, info governance, org & roles, info life cycle, and enables infrastructure
    21. Example: launch big data project. If fails b/c of data quality, will blame project not bad data. No obvious strategy.
    22. Command and control the information strategy, vision. Create plan, identify key people in the org to execute
    23. CDO creates the organization that executes data vision for lifecycle of data - from ETL to quality to archiving
    24. Main job #1 = drive effective information governance. Take into account policies, regulations, define standards
    25. Why will this role grow? Because digital business is confusing. Someone needs to craft information strategy
    26. When it comes to information risk, 75% of companies are still focused on risk and compliance #BusinessCase
    27. How much time is spent fixing data quality issues, dealing w risk? If this exists in your org, you need strategy!
    28. When it comes to information opportunity, only 50% are using information to enhance product and service offerings
    29. When it comes to information opportunity, only 10% are using information to boost product development.
    30. Chief Data Officer can tackle all of this - it's an innovation role to harness power of data within the org
    31. CDO Business Case = Information-driven decision making > data monetization > digital remastering
    32. Desire to deliver competitive advantage through the collection, analysis and use of info. (1/2)
    33. CDO thinks about risk of damage from loss/inappropriate use of info. Marketing/Sales/etc aren't doing that. (2/2)

How to Leverage Infonomics to Measuring the Value of Information as a Corporate Asset

  1. Presenter:  
    1. Douglas Laney, Research VP, Gartner
  2. Summary:
    IT and business leaders often talk about information as a key strategic asset, but fail to manage or measure it as one. This session provides the techniques for assessing the value of information assets. 
    1. Should information be considered an enterprise asset? 
    2. How can you measure and improve information value and infocentricity? 
    3. How can you become an infocentric organization?
  3. My notes (tweets):
    1. Infonomics term came out of September 2001 attack related to claims for loss of data assets.
    2. Insurance won't cover claims for data as an asset. Accounting won't allow capitalizing data on balance sheet
    3. Courts are split that data is physical. It's on a disk but some say "electrons have no mass" so data isn't "real"
    4. What is an asset? What is information?
    5. Information won't be seen on the balance sheet of Facebook, Twitter, etc.
    6. What is an asset? Can be owned, can be sold, generates potential future economic benefit. Data fits all of these
      1. Information isn't a consumable, but what about brands/patents/trademarks? Same deal, those are assets.
        1. "Tobin's q" = ratio of market value to tangible assets. Avg = 1.1. Info-centric cos = 2.4, Info-product cos = 4.7
      2. Information is an actual asset (even if not recognized as such)
      3. Information has potential, probable, and realized value
      4. Information's value can be quantified
      5. Information can be reported on internally
      6. Information's value should be used to justify business decisions
      7. Information should be managed and leveraged as an asset
    7. Why? Can't manage what you don't measure. Project & confirm ROI of Information Management initiatives. (1/2)
    8. Using info as a currency. Assessing risk that contracts/insurers deny info is property. Productizing info. (2/2)
    9. Financial analysts will start to include data assets as part of a company's valuation.
    10. Measure gap between potential, probable, and realized value of information
    11. @Doug_Laney looked at Michael Lewis's "The Real Value of Everything" - guess what wasn't in there? Information!
    12. Value can be measured: Relative value (internal measure), Financial value (cost/benefit, opportunity)
    13. How much better would salespeople be able to perform if they had access to competitive data?
    14. Think of ways that the information you have, or do NOT have, figures into the "value" you can assign it
    15. What was data acquisition cost vs benefit gained from winning a deal/customer?
    16. What kind of revenue can we generate from the information minus (administrative + acquisition + application costs)?
    17. How about the market value of the data? Income generated by "selling" it (technically it's just licensing)
    18. Asset management: Financial, material, human, intel prop, information (can't own people but can manage as assets)
    19. Think about mgmt of physical assets. Can you apply those principles / techniques / learnings to asset mgmt?
    20. Communicate your intention to measure value of assets. Collaborate with peers to find ways to measure & assign
    21. Can Information be a liability? Not from accounting perspective, but from a risk perspective definitely yes.
    22. Any tools to manage data assets? Profilers can count records/field lengths or can use data quality tools to measure

Why You Need a Single View of the Customer - and Why MDM Is Only Part of the Solution

  1. Presenter:  
    1. Bill O'Kane, Research Director, Gartner
  2. Summary:
    MDM is a critical part of constructing the coveted "360° view" of your customer, but MDM is not the entire solution. Learn what the other components of the 360° view are, and how to provide them along with MDM. 
    1. How does MDM enable the 360° view of the customer? 
    2. What non-MDM components are required for the entire solution? 
    3. What are best practices for managing implementation of the entire solution?
  3. My notes (tweets):
    1. "Single View of Customer" is really interesting to me from multiple angles: MDM, EDH (Hadoop), Tableau. My current life revolves around it.
    2. Agile data gov req'd when data is outside your control. Someone else's master data/data from apps will be slippery
    3. Yet another mention of @mdmcentral's "Three Rings of Information Governance." Must read!
    4. MDM is not for all data. MDM is only about data that is worth managing and can be managed.
    5. Master Data is its own category (vs Content vs Analytic vs Transactions vs Social Data vs Dark Data vs Event Data)
    6. Teams w/in governance org work together to support different data categories. Vision, Strategy, Metrics unified across all.
    7. One data category that isn't on the prev list is Metadata. It's "meta" because it spans all, relates to all
    8. Mature the MDM program to the EIM scope - apply 7 bldg blocks framework to everything practical w/in the data sphere
    9. Just because EIM don't need a lot of transformation doesn't mean they aren't worth applying rigor to
    10. Become comfortable with concepts of "trust vs truth" and "fitness for purpose"
    11. .@BillOKane "Trust vs truth: you can trust something that's not true if you understand how not true it is"
    12. "Data Integration Hub" can be used to combine biz data next to MDM Hub. Don't exceed capacity or use case profile
    13. Customer Service & Sales are the two orgs most likely to want Single View of Customer
    14. A large % of customer MDM is not secure at the data level. Usually trusted-user set up - @BillOKane expects this to change

SAS: Three Key Data Management Practices for Analytics

  1. Presenter:  
    1. Dan Soceanu, Senior Solutions Architect, SAS
  2. Summary:
    1. Recent advances in business analytics give today’s enterprises more power than ever to make better decisions. However, the most common pitfall is not the analytical tools, it’s the data management required to successfully use advanced analytics. This session will explore three key data management practices that are essential for every industry and within all data landscapes. 
  3. My notes (tweets):
    1. Context is critical. Data from the IT perspective is very different than from business side
    2. Information is data in context. Knowledge is information in context. Wisdom is knowledge in context.
    3. Wisdom allows us to make a decision. Data > information > knowledge all lead up to that. Can't decide based solely on data.
    4. Categories of context: Computing, user, physical, and time. Human brain superior to today's computers bc context dimensionality.
    5. Jeopardy Champs vs IBM Watson was sleight-of-hand. Watson had questions before hand. The qs it got wrong bc of context
    6. Tools / techniques for integrating data were designed for data warehouses, not for ad hoc, agile, big data
    7. Multi-billion dollar corp posed problem to contractor: tell me how many employees we have? Their daily stat was +/- 40K people!
    8. Analytics isn't worth anything if your data can't be managed properly
    9. Biz metadata vs tech metadata. Biz = biz rules, definitions, lineage. Tech = source/target systems, schema, dependencies
    10. Analytics lifecycle with roles (context)
    11. Analytics maturity scale
    12. Studies show trying to improve data quality after it's landed in repository is 10x more expensive & difficult to do
    13. Holistic data management, given context and systems alignment
    14. Analysis needs context: policies for sharing data across enterprise. Define data meaning/labels. Quantify business impact.
    15. Put the data into action (decision making is an action). Data doesn't exist in a vacuum - have to give it an operational purpose
    16. SAS has adapted #SixSigma to drive their analytics practice
    17. Summarizing it all - the recommended structure for the analysis ecosystem

How to Identify and Mitigate Information Risk

  1. Presenter:  
    Saul Judah, Research Director, Gartner 
  2. Summary:
    Risk management is a critical component of your EIM program, but how do you go about it? Using a case study, this session will help you understand how to identify and mitigate information risk. 
    1. What is information risk and why should I care? 
    2. How can I identify the information risk relevant to my EIM program? 
    3. What should I do to address information risk in my enterprise?
  3. My notes (tweets):
    1. Key action: how to act now to identify risk using EIM and risk assessment frameworks.
    2. Information risk is everywhere. Understand the difference between information risk & information program risk
    3. Data quality - doing this so we know the data is good enough to trust. Poor quality == poor results
    4. Why are we doing MDM? Apart from single view of customer/product/etc, do it so we can take advantage of the data's value
    5. Evaluating all information risks in a matrix is a perfectly valid approach. But it may not be enough
    6. Ideally, you should explore, access, and express your information risks. What does the landscape really look like?
    7. Explore information risks by talking to your people: 1:1 interviews, surveys, brainstorming, scenario planning, etc.
    8. Explore information risks by evaluating systems: threat mitigation plan, security scans, etc.
    9. Assess result of exploration across people, process, information, infrastructure for info risks. Rank/model/calculate outcomes
    10. Once you know what your information risks are through the explore/asses/express framework, you can create a mitigation plan
    11. Use your information strategy to rank priority for mitigating information risks. Incorporate into your plan
    12. Now that you've identified the full lifecycle for information risk mitigation, review your overall plan
    13. None of these on their own will help if your culture doesn't acknowledge the risk or the mitigation approaches. Will be unique
    14. Final outcome is a mix of information risk strategies based on tolerance to risk outcomes across biz units, culture, regulation
    15. Will improve risk maturity. Accepting risk based on "risk appetite" (think deductibles on your insurance). It's about trade-offs
    16. Information risk evaluation/action plan outcome may look like this

How to Make Progress on Data Quality Improvement

  1. Presenter:  
    Ted Friedman, VP Distinguished Analyst, Gartner 
  2. Summary:
    Data quality is a critical starting point for various types of information governance initiatives, and poor-quality data continues to degrade the value of many business initiatives. Apply the key principles that all organizations need to internalize in order to make progress on this issue. 
    1. Why does a focus on data quality help to bolster the business case for information governance? 
    2. How can you apply current and emerging best practices to make progress?
    3. Where can you leverage data quality technology for best impact?
  3. My notes (tweets):
    1. Five ideas that put a different spin on DQI efforts.
    2. How many people have performed rigorous measurement on the cost of poor data quality?
    3. Need to understand how much poor data quality costs in terms of headcount, lost time, customer loss (!?), etc.
    4. Gartner does a yearly study on data quality. Asked 385 co.s for self-assessment of cost. ~50% estimated $1M or more per year
    5. ~35% don't even know how much it costs. Also, @ted_friedman believes the costs are underestimated for those that report a cost.
    6. Avg estimated annual cost of poor data quality = $13.3 million (USD).
    7. Five things you must do to make progress: 
      1. Use the right words to describe the importance of poor quality/quality improvement
      2. Scoping is key. Set a meaningful scope - target data that is business- or mission-critical and improve that to see gains
      3. Get the right roles and skills in place. Stewardship FTW.
      4. Define metrics and measure your quality.
      5. Apply tools when and where they can add value
    8. #1.) Use the right words
      1. "Quality is bad. Make it better to get value" = NO! What does that mean? "Key initiatives are hindered. Fix X to gain Y" = YES!
      2. "We need a full-time people to fix this" = NO "We need commitment from all & a few critical formal roles to fix this" = YES
      3. "Buy tools to clean data" = NO! "Tools are a part of solution but don't deliver value alone" = YES!
      4. "We will build quality dashboard!" = NO! "We will measure and monitor DQ levels to quantify financial/biz impact" = YES!
    9. #2.) Scope is key. Don't spend time/money to fix random data sets. Determine if data is in scope for DQ efforts
      1. Determine scope of DQ effort. Ask of data: Does it influence critical processes? Is it visible to external customer/regulator?
      2. Ask of data: Is it subject to legal/regulatory mandate? Can it be leveraged for competitive advantage or sold outright?
    10. #3.) Roles/skills - Important roles for data quality: data stewards (across org), high-level sponsorship, information architect, analyst.
      1. If data influences critical biz processes, then need to have roles attached to those processes & stewardship for the data
    11. #4.) Define metrics, measure them, and expose the results to people who care. Pick the relevant data that works for your env.
      1. Like "choosing your words wisely," make the data quality metrics meaningful to everyone - don't pick obtuse data points.
    12. #5.) DQ tools profile, measure, visualize, monitor, parse/standardize. They can't interpret metrics, discern rules/cleanse complexity
      1. Data quality is easy when you have a standard field to compare. But how do you measure quality of content (a video, a Word doc)?
    13. Proactive + monitoring + automation + justification/scope + info governance pgm + continuous improvement + visibility to metrics

How to Explain to Your Executive Team the Value of EIM in Digital Business

  1. Presenter:  
    1. Jim McGittigan, Research VP, Gartner
  2. Summary:
    1. Engaging the C-suite on EIM is challenging. This session explores what is important to the C-suite and the board and explores ways and methods to relate EIM to their needs. 
      1. What are the priorities of the C-suite and board? 
      2. How does EIM relate to C-suite goals and objectives? 
      3. How can you turn data into money? 


Saturday, March 21, 2015

Debfrief: SFBATUG Event at Twitter HQ March 19, 2015


I attended the San Francisco Bay Area Tableau User Group (SFBATUG) event held at Twitter HQ in downtown San Francisco on March 19, 2015.

The session offered four "lightning" presentations from Tableau legends/Zen Masters John Mathis (Slalom Consulting), Daniel Seisun (Twitter), Anya A'Hearn (Datablick), and Allan Walker (Slalom Consulting). Ashley Ohmann (K-Force) officiated the event.

As this tweet shows, it was a packed house:


First up was John Mathis. John showed the audience how to use Tableau to present complex data generated in R. The process involved calculating the values in R on a local R server, then saving the result set as an RDA file. Directing Tableau to the data set in that R file means one can use Tableau's superior graphical engine and viz tools to make the data come to life for an audience already using Tableau.

Of course, if you are an R wiz, used to delivering results from R to your constituents, adding Tableau into the mix isn't a required step. As with any tool chain, consider the consumer of the data and the mechanism that makes receiving - and understanding - the data most convenient.


Next, Daniel Seisun gave a primer on the use of Tableau at Twitter. Daniel's mission has been to move from an in-house BI tool to Tableau while maintaining the flexibility and DIY company culture that the in-house tool provided.

He described to the crowd that his team - of two people, including himself - are responsible for Tableau's adoption and administration across the entire company. Any employee who wants to play with Tableau gets their own virtualized instance to play with. This "sandbox" approach ensures lower administrative overhead for Daniel's team while providing an ad hoc environment for his customers.

Finally, he walked the audience through a series of Python scripts that his team uses to clean out old viz, ensure that users are actually using the tool (and if not reclaim licenses for other new users), and keep the instances performing well.


Daniel suggested that these scripts could be open-sourced. Allan Walker responded by setting up a Tableau git hub moments later. Visit it here and contribute today!


Anya A'Hearn presented next. She told the audience how to find Love on Twitter. No, really.

Anya began showing attendees how to create custom maps (map layers) with MapBox. With custom maps in hand, loaded into Tableau, she then combined layers with geo-coded data from tweets containing #love to create a viz that allows a user to interact with tweets in their area.

Finally, she presented a web-based interface that allowed speakable commands to toggle map layers on and off by person, location, and sentiment.


The final presenter, Allan Walker, took customization to a whole new level. He focused on the use of Tableau's JavaScript API to take the viz out of the server environment and give it new life in a new context on the web.

Allan showed the audience several websites with this in action. The first, a simple demo where the viz could be moved around a webpage. He noted that by combining multiple dashboards on a single webpage, the end user could interact with the data in all new ways.

Next he showed how Tableau data could be combined with live map data through JavaScript (hereafter "JS") handlers that send parameters in Tableau to the map to produce customized walk-throughs of San Francisco International Airport.

Taking it out of JS and into WebGL, Allan demo'ed an interactive globe object showing earthquakes around the world. The lucky few attendees with an Android device got an extra "jolt" as this viz/web combo had a haptic feedback program embedded that gave viewers a vibration when interacting with quakes on the map.

His final demo showed how the viewer's own body could be read by the device camera via JS and motion conveyed to an app. Allan's key message was that we need to think outside of our static web framework and consider how we can make the data come alive to any viewer of any ability level. This was a poignant comment that really resonated, and inspired a lot of people in attendance.

Inspired as they were, many have little-to-no JS experience and asked how to get started. Allan's recommendation is to find the many test projects already documented and posted on the web, and use them as a learning tool. Of course, people may wish to undertake a JS primer first. Many such options are available online.


The event was a lot of fun. Kudos to great programming by Anya A'Hearn, excellent work in keeping things rolling by Ashley Ohmann, and of course big thanks to Twitter for providing the space, the food, and the event coordination!

Friday, February 6, 2015

Tableau Desktop and Server Resources

As a Tableau Desktop & Server user, I'm always  on the lookout for helpful resources.
The list below compiles items I found on my own, with a few suggestions from my company's account rep at Tableau:
If you have a favorite resource I didn't list here, please add it in the comment so I can incorporate into my list.