Карта сайта
Версия для печати

Как подружить бизнес и ИТ?

25 июля 2013 Отличий между представителями департаментов бизнеса и ИТ настолько много, что порой даже собственник бизнеса может принять их за сотрудников разных компаний. Вдобавок к общепринятым стереотипам (первые считаются экстравертами, в то время как вторые - интровертами), самый типичный пример конфликта - это проблема несвоевременной поддержки бизнеса сотрудниками ИТ и выполнение технических работ, связанных с BI-системами, которые мешают либо ограничивают работу бизнес-пользователей. И все же существует простой способ налаживания взаимопонимания. Какой? Читайте в нашей статье! (Материал опубликован на английском языке)
Advancements in analytics, however, may well prove the proverbial middle ground that provides the crucial truce, so IT can happily supply business users with access to meaningful data to decrease time to use and inform decision-making.

The importance of analytics in providing this crucial bridge was discussed during an Enterprise Data World 2013 session entitled “Experts on Analytics” hosted by Neil Raden of Hired Brains which featured Brian Sletten of Bosatsu Consulting, Inc., Leonard Asuncion of Vertica, and Brian Crook of Verve Mobile. Critical factors defined in ameliorating this relationship include the deployment of Semantics technology, the role of Data Scientists, and the effectiveness of Data Visualization tools. Sletten explained that:

“A fundamental part of the problem of how we make choices is the perception that IT and business are not working for the same organization, that IT is not supporting the business and that the business isn’t making decisions in terms of IT. Collectively, we all have to stop thinking that this is a separate function. IT supports the business, and not the other way around.”

Keeper of the Peace: The Data Scientist

In many ways, analytics is at the forefront of conventional Data Management, with a number of drivers including NoSQL and Big Data technologies increasing the quantities of unstructured data. The gradual shift from structured, relational database information to that gained from the Internet, social media, sensor technologies, and other sources has also heralded a shift from Business Intelligence tools providing historical analysis to predictive analytics.

Data Scientists are largely charged with cleaning and deriving structure from what is typically regarded as unstructured (which really means non-relational) data. Their job is to translate data so that business and C-level users can create action from it. They are the frontline of IT-associated techies who can derive meaning and value from data to provide to business users, and play an integral role in assuaging the divide between these two domains. Their actions require less of an ability for end-users to be cognizant of mathematical-based statistics concepts than it does for Data Scientists to model data in a way that laymen can understand. According to Crook:  “It’s hard to define a Data Scientist. Not only do they have to have a great understanding of math, but they have to understand the business domain. It’s easy to talk about Data Scientists in terms of what they do, but they’re not easy to talk about in terms of the right kind of personality, background and skill sets. They’re going to be different for every organization.”

Although there is a definite increase in the number of formal education institutions offering collegiate degrees in aspects of Data Science, there are actually varieties of different methods and curriculum incorporated into Data Scientist training programs. According to William Cleveland is his seminal document regarding Data Science, this discipline should incorporate:
  • Multidisciplinary investigations
  • Models and methods of data
  • Computing with data
  • Pedagogy
  • Tool Evaluation
  • Theory
Since the goal is to help technical users demonstrate the value of data to non-technical users, a number of different approaches apply. Some organizations have their own separate training for individuals hired for this position; others implement continuing education requirements to assist in this endeavor. Several collegiate programs have various points of focus outside of mere science to emphasize business objectives.  A number of audience participants discussed the fact that social perceptions of math and statistics should be reconfigured so that such fields have less of a stigma and are more readily accessible to all data users. Regardless, the Data Scientist has the principle duty of mastering a number of different disciplines – aspects of data, business, statistics, social intelligence, communication – to effectively relay information about data to non-technical users.

Visualization: Making Data Easy

There are also a number of technological advancements in analytics that can bridge the rift between IT and business. Self-service BI tools – many of which incorporate predictive analytics – are certainly on the rise and are aided considerably by advancements in Data Visualization. Visualization tools enable users to obtain visual representations of trends in data, often by simply inputting a relevant set of numbers.

Data visualization tools (such as Tableau) are tailor made for non-technical users, since they show data that otherwise might take considerably longer to report with conventional BI tools.

Data discovery tools can do much of the same thing and are integral to elucidating salient factors related to data. These analytics applications generally require IT personnel to configure them to suit the particular focus of end users and allow them to expediently apply analytics to their job functions by self-querying data at their leisure. Most of these tools utilize pre-defined models and algorithms which offer likely hypotheses that are best fully explored by IT before moving into production. Still, the proper configuration of visualization tools is one of the most tangible wins for IT departments, and a definite way to ingratiate itself to business users.

Sletten commented on the fact that the demand for Data Science programs as a formal discipline coincided with the emergence of useful visualization applications:  “It’s no coincidence that this eruption of Data Science among trendy, hipster Web 2.0, 3.0 companies coincides also with the explosion of Data Visualization frameworks. I think it’s also not a coincidence that the most widely used or most successful one, D3.js by Mike Bostok who came out of Stanford Visualization Group is also data driven documents. It’s not a florid, overly complex presentation; it just lets the data speak for itself.”

Semantics Solutions

Another means by which IT can provide substantial support to business users is through Semantics technology. Semantics solutions are of inestimable value for providing analytics because they reduce all data – regardless of structure or source – to a shared structure, a triple, in which it can be analyzed in terms of its actual meaning as opposed to its representation in a particular model. Triples are simple sentences that describe data in terms of what it is, what it does, and what it does things to that are crucial for federating and aggregating data.

Semantics solutions may be one of the most useful tools for Data Scientists to employ to render the meaning of data in a method that is independent of conventional modeling or reporting tools. The most commonly used Semantics technologies include a modeling framework, RDF, its principle query language SPARQL, and OWL, a language that describes data relationships and attributes. Semantics has the potential to significantly reduce the complexity associated with determining meaning from data – especially multi-sourced and Big Data – to enable business users access to varying data types and to attributes of data that are not otherwise readily discernible.  Raiden commented on the viability of Semantics for analytics: “We’ve been working under the assumption all this time that getting data, like in a data warehouse, really prepares us for analytics. But without the application of Semantic technology, we’re missing a lot of information that’s either there or can be deduced without our working really hard”.

Happily Ever After?

While business users and IT departments may never fully reconcile all of their differences, a number of them can be ameliorated through the successful application of analytics. Analytics technologies help remind these respective domains of their common interest, and are the critical means by which IT can thrive in its role of assisting the business. As Semantic Web technologies slowly make their way into the enterprise, end users will be able to discern more significant relationships between data. Visualization tools (configured by IT) give business users more autonomy in their perception of data’s significance. Ultimately, it is up to a new generation of Data Scientists to incorporate a variety of skills to effectively translate lifeless data to users who can transform it into productive action.


Source:  dataversity.net