Reference : Below contents are from a stanford white
paper. I am publishing the content here to spread a word about the
article and content. The content can be accessed from the URL given at
the end of the article here..
Thanks to the authors and Stanford for the advanced topics and the emerging trends in the area of Business Analytics.
Business Analytics:
The field of business analytics has improved significantly over the last few years, providing
business users with better insights, particularly from operational data stored in transactional
systems. As an illustrative example, analysis of e-commerce data has recently come to be
considered a killer-app for data mining. The data sets created by integrating clickstream
records generated by web sites with demographic and other behavioral data dwarf, in size and
complexity, the largest data warehouses of a few years ago, creating massive databases that
require a mix of automated analysis techniques and human effort in order to provide business
users with critical insight about the activity on the site and the
characteristics of the site’s visitors and customers. With many millions
of clickstream records being generated on a daily basis and aggregated
to records with hundreds of attributes, there is a clear need for
automated techniques to find patterns in the data. In this paper we
discuss the technology and enterprise-adoptiontrends in the area of
business analytics.
The key consumer of these analytics is the
business user, a person whose job is not directly
related to analytics per-se (e.g., a merchandiser, marketer,
salesperson), but who typically must use analytical tools to improve the
results of a business process along one or more dimensions (e.g.,
profit, time to market). Fortunately, data mining1, analytic
applications, and business intelligence systems are now being better
integrated with transactional systems creating a closed loop between
operations and analyses that allows data to be analyzed faster and the
analysis results to be quickly reflected in business actions. Mined
information is being deployed to a broader business audience, which is
taking advantage of business analytics in everyday activities. Analytics
are now regularly used in multiple areas, including sales, marketing,
supply chain optimization, and fraud detection.
The Business Users and their Challenges
Despite these advances in analytic systems, it continues to be the case that the business user,
while an expert in his area, is unlikely also to be an expert in data analysis and statistics. To
make decisions based on the data enterprises collect, the business
user must either rely on a data analyst to extract information from the
data, or employ analytic applications that blend data analysis
technologies with task-specific knowledge. In the first case, the
business user must impart domain knowledge to the analyst, then wait
while the analyst organizes the data, analyzes it, and communicates back
the results. These results typically raise further questions and hence
several iterations are necessary before the business user can start
acting on the analysis. In the second case, analytic applications must
not only incorporate a variety of data mining techniques, but also
provide recommendations to the business user of how to best analyze data
and present the extracted information. Business users are expected to
better utilize the extracted information and improve performance along
multiple metrics. Unfortunately, the gap between the relevant analytics
and the critical needs of the intended business users still remains
significant. The following challenges highlight characteristics of this
gap:
1. The time to perform the overall cycle of collecting, analyzing, and acting on enterprise
data must be reduced. While business constraints may impose limits on reducing the
overall cycle time, business users want to be empowered and rely less on other people to
help with these tasks.
2. Within this cycle, the time and analytic expertise necessary to analyze data must be
reduced.
3. Clear business goals and metrics must be defined. In the past, unrealistic expectations
about data mining “magic” led to misguided efforts without clear goals and metrics.
4. Data collection efforts must have clear goals. Once metrics are identified, organizations
must strive to collect the appropriate data and transform it. In many situations, data
analysis is often an afterthought, restricting the possible value of any analysis.
5. Analysis results must be distributed to a wide audience. Most analysis tools are designed
for quantitative analysts, not for the broader base of business users who need the output
to be translated into language and visualizations that are appropriate for the business
needs.
6. Data must be integrated from multiple sources. The extract-transform-load (ETL)
process is typically complex and its cost and difficulty are usually underestimated.
