Theses and Main Ideas

  • Natural language humanizes Data Analytics, making it easier to be widely adopted.
    • Natural Language Processing (NLP) helps computers understand the meaning of human language.
    • As natural language evolves within the Business Intelligence (BI) industry, the barriers to Analytics adoption will fall, transforming workplaces into data-driven, self-service operations.
  • Acelerated cloud data migration fuels modern BI adoption.
    • Data gravity means that applications and services are pulled in the direction where data resides.
    • Data moves to the cloud at an accelerated rate and Analytics naturally follows.
    • This increases agility and recasts the possibilities of what you can do with BI and Analytics.
  • Actionable Analytics brings data into context, where it’s needed:
    • BI platform vendors are offering capabilities like Mobile Analytics and Embedded Analytics.
  • It is the integration of BI tools –that is, Analytics contentsinto business process applications.
  • To help the user work smarter, it provides relevant information and tools designed for the task at hand.
  • The Analytics functionality is embedded in:
    • Dashboards
    • Data visualisation tools
    • Visual workflows
    • Interactive as well as mobile reports
    • Self-service Analytics, etc.
  • It studies of the way mobile users access the Internet and specific websites.
  • It covers statistical information such as the number of people using a particular operating system or handset.
  • It can also cover behavioral information: how long mobile users spend visiting a site compared to people using a desktop computer.
  • Data Science teams will use Explainable Models that reveal how they are constructed.
  • Business users need better understanding of their data environments and assets to leverage them effectively.
  • Organisations that rely on AI and Machine Learning for data-driven decision-making, show a rise in human hesitation about the trustworthiness of model-driven recommendations.
  • There is a lack of Machine Learning applications that provide a transparent way of seeing the algorithms or logic behind decisions and recommendations.
  • Managers are not going to take decissions if they don’t understand the causality.
  • Explainable Models also clearly contribute to making it easier for firms to work with the Agile operative framework.
  • Data projects need to be implemented quickly and easily in less time, to get short term results. That is
  • Agile Data is the term for the data companies need right now.
  • The Data Layer provides the speed, reliability, scalability and agility needed to thrive in today’s complex business landscape.
  • The Data Layer is needed to ensure compliance:
    • Organizations can get a 360-degree view of every interaction so they can respond to compliance requirements with ease:
      • Examples:
        • HIPAA (Health Insurance Portability Act)
        • SOX (Sarbanes-Oxley)
        • GDPR (General Data Protection Regulation).
      • How an enterprise Data Layer drives innovation:
        • Increasing productivity: Data Layer offers data autonomy and complete control of your data, wherever it resides.
        • Ensuring compliance: organizations can get a 360º view of every interaction, so they can respond easier to compliance requirements –e.g.: HIPAA (Health Insurance Portability Act), SOX (Sarbanes-Oxley), GDPR (General Data Protection Regulation).
        • Delivering personalized experiences: technology can uncover relationships between disparate data sets, ending up with more sales and resources to support innovation. The tools and data employees need are always readily available, making their work more efficient.
      • Increasing business agility: as market dynamics and businesses shift, companies have the real-time insights they need to make the right decisions quickly.
      • Driving competitive advantage: when enterprises’ need change, they derive new solutions, to get ahead of their competitors.
      • Enterprises are flexible to store, run and process data, so they can move faster and innovate quicker
  • Data quality impacts firm performance directly and indirectly, so it must be part of a BDA (Big Data Analytics) and IoT (Internet of Things) strategy in order to create business value for organizations.
    • There is a BDA and IoT business value model that examines the influence of data quality on the functioning of these capabilitires, mediated by business process sophistication. The model also assesses the impact of a BDA and IoT on competitive advantage.
    • Quality over quantity: the quality of data and the ability to protect it is more important than the quantity.
    • The risk from unused data must be eliminated.
    • Getting a handle on data lineage: correct data lineage is a full understanding of the data, its transformational nature, its associations and its lifecycle across the data estate and over time.
    • Data must be of good quality, controllable and consistent, to ensure that it can be relied upon to feed these technologies appropriately, to generate business value and inform decisions.
    • All Analytic tools and AI models require quality data to make accurate predictions or to provide the best insights.
    • Executives from a wide range of companies, government agencies and departments, tried to represent the percentage of data correctly created (their Data Quality (DQ) Score) using the Friday Afternoon Measurement (FAM) method *.

    ——————————————–

    * Nota: The FAM is a 4-step method to determine the quality level of your data (See Recommendations for a full account of FAM).

Spanish bank BBVA (use case)

It monetizes insights: to be sold to anyone interested

  • Geoblink: first company to use the BBVA API Paystats
  • It enabled it to complete its offer of geolocated data in Spain, including the option to access info on:
    • The hourly patterns of comsuption
    • The average ticket for a comercial category.

Or where consummers to a retail area are coming from.

  • At the beginning, each firm reinvented the wheel.
  • Until we realized we all wanted:
    • The same data (mostly).
    • To be treated the same way.
    • And be used similarly.
    • What they sell is an API to be called for raw or processed data

Roles & Responsabilities:

  • Many organizations find identification of cross-functional roles the most difficult hurdle when developing the roles in their Data Governance program.
  • Finding the right people to fill the roles associated with decision making for a specific subject matter of data is tricky.
  • Key role: Data Protection Officer, who modifies business practices and, above all, knows what personal data is stored, who is using it, how and for what purposes.
  • The role of CIOs and CDOs:
    • As Gartner defines them, CIOs (Chief Information Officer) and CDOs (Chief Data Officer) must support each other in delivering business, and neither can succeed alone.
    • Both Chief Data Officers (CDOs) and Chief Information Officers (CIOs) may work in technology departments and must work together to pave the way for a data culture within the company.
    • Chief Data Officers (CDOs) are at the forefront of the fight for data Analysis.
    • Chief Information Officers (CIOs), on the other hand, oversee the technological development of a company (IDEA). Read more 

How Data Governance helps business transformation:

  • Data Governance breaks down data silos from disparate systems across the enterprise, and establishes a set of processes, standards and policies to make the data consumable enterprise-wide.
  • Having systems in place enables users to understand data in business terms, while also establishing relationships and associations between data sets.
  • Data Governance needs to emerge as a separate area within companies, in order to:
    • Better maintain their data
    • Facilitate the use of data.
    • Exercise control over processes and methods employed by all data users across the business.
  • The CIO (Chief Information Officer) needs to collaborate with the CDO (Chief Data Officert) to manage a data strategy underpinned by governance that delivers the right data into the right hands to drive business transformation.

Data Governance strategies:

  • Governance efforts should be tightly linked to digital efforts.
  • Creating a strategy around data is pivotal to a successful Digital Transformation (DT).
  • It’s imperative to employ Data Governance strategies before organizations can fully embrace the full potential of these technologies (cloud, IoT, AI and Machine Learning).
  • Using data to drive insight is not easily understood, notwithstanding new data privacy regulations and issues with data ownership & algorithmic bias.
  • To be used for faster decision making, users must trust data (as the corporate asset it is).
  • Data must be correct and unbiased for technology to have the right impact: flawed data only slows Digital Transformation.
  • Collaboration is key to successful data use across an organization.
  • A Data Gobernance strategy sets criteria for:
    • Traceability of data
    • What data sources we´ll use and why
    • Data quality standards
    • Who are the data owners at each phase of the intake, storage and processing sequence?
    • Who can use the data? Everybody or just a few select ones?
    • Which ought to be the tech criteria?
      • Open source?
      • Single provider?
      • On-the-shelf?
      • Self made?
    • Data management style:
      • Empowered and self-governing teams?
      • Or the traditional project manager command style?
  • Companies find that skills shortages are hindering the potential of Data Analytics:
    • These shortages translate into a lack of expertise in integrating multiple datasets…

And into a failure of understanding in order to deploy the right analysis techniques.

  • Four key pillars rule the lifecycle of Analytics projects:
  • Data acquisition
  • Processing
  • Surfacing
  • Actioning on the data.
  • (Each contributes to a significant part of the value chain of Analytics).

    • Data acquisition: a wide range of tasks, systems and technology knowledge needed to be effective in acquiring the required data.
    • Data processing transforms raw data and refines it into informative data.
    • Informative data needs to be surfaced in an effective manner to be meaningful.
    • Actioning: Analytics needs to provide actionable insights (“Analytics without action is just research”)
      • We need to focus on the last stage of the value chain of Analytics: Action.
      • Analytics has three separate subdomains: Descriptive, Predictive and Prescriptive Analytics
      • The latter provides a prescription for action.