In my previous post I had described what business intelligence
is and how BI analysis is different from normal IT business analysis. In this article I will try to explain the
skillsets which are required for being a successful BI analyst.
Technical Skills
BI analysis requires knowledge of data warehousing and using
analytical tools to analyze data. There are many BI tools available in the market
and major ones are IBM Cognos, SAP BusinessObjects, Oracle Hyperion, Microsoft
SSAS and SSIS. However I believe if you gain command over SQL and Excel (VBA)
for BI reporting, then you can use any of the above tools quite easily and your
learning curve will be very smooth. You can easily download trial version of SQL
Server 2008 onwards and Microsoft excel to work on them. There are myriads of
tutorials available on internet to learn sql and excel, which can be used. On top of that, I
will give you practical BI cases which use SQL and Excel to analyze a business
problem in future posts.
More sophisticated BI
employs data mining, business modeling and pattern recognition which use
statistical modeling softwares such as SAS, SPSS and R. Knowledge of statistics
and solving practical problems using any one of the tools can help you learn it . However it is better to join professional classes to venture in this
field.
Data Extraction and
Manipulation
For analyzing the data we need to extract it first. The data
lies in the database from where it has to be retrieved. Knowledge of SQL is
very important as it is used to extract the data. Generally the information
required to do the analysis is present over 2-3 tables. Hence, knowledge of SQL joins (inner join, left join, right join) is
a must. The extracted raw data which is used
for analysis is called as ‘dataset’. E.g. sales data of 100 shampoo bottles,
containing product number, cost, margin, geography, can be termed as a dataset.
Business Intelligence
without statistics
Normally BI reporting uses SQL or SAS to transform the
retrieved data. Transformation such as cleaning, sorting, merging tables, aggregating, grouping and transposing are done on
the data to obtain required results. Then final output is represented in
graphical format using excel charts and graphs. They are made dynamic by using
pivot tables to represent the information.
Pivot filters can be checked and unchecked to roll up or drill down
information. E.g. sales of the shampoo have been aggregated on city level (e.g.
Mumbai, Delhi and Bangalore) and for 3 months. You can manipulate filters to gain insights on
the sale of the shampoo geographically and monthly.
Business Intelligence
with statistics
Statistics is used in
BI for creation of business model, so that ‘what if analysis’ (running the
simulation for various conditions to reach optimized results) can be
performed and forecasting can be done. Data mining is also very commonly used term in
the field of BI, which uses statistics to recognize patterns from large datasets.
While using statistics in Business intelligence, following topics come into use
regularly: regression analysis, correlation, factor analysis, business
modeling, forecasting, hypothesis testing, clustering etc. Hence, working
knowledge of statistics is very important.
Generally SAS, SPSS and R are the common statistical tools,
which are used to do statistical modeling. These are very powerful analysis
tools which can handle large amount of data, running into millions of rows.
Sometimes Excel is also used to do statistical analysis when dataset is relatively small.
Soft skills
Most sought after soft skills for
BI analysts are good communication skills, personable and presence of mind. You
have to Interact with business managers for gathering data, discuss analysis to
be done and what are the required results management is looking at. There is
lot of ad-hoc analysis too. Your analysis will influence their decisions; hence
good rapport and communication skills are very important. These traits help you
to understand correct requirement and successful synergy between you and the
business managers.
BI analysts need to have an
eye for detail. There have been lots of times when a wrong decimal point in
numbers can make your reports go haywire.Hence, Quality Check (QC) is very important
aspect of BI reporting.
One small example, let’s take 3 shampoos A,B and C. We have to
present their sales in form of percentages to total sales of the 3 shampoos. During QC, we should
check that sum of these percentage add up to 100. Similarly, you should always have checking points in the report which ensure sanity of the reports.
Domain Knowledge
Last but not the least, domain knowledge or in other words, knowledge of the industry for which you are working is very important. It can be banking, finance, supply chain, sales, marketing, HR, I mean anything.
I believe anyone can learn the tools and statistics for generating insights from the data. But the insights given by a person who has experience in that industry and has knowledge of these analytics tools will deliver even more matured insights. Obviously, a person with more information and experience, can connect the dots in a better way and give valuable insights. Hence, always try to know about the business of your company in depth, latest news, future strategy which is very important for a business intelligence analyst.
