Data computing & analytics software (DCAS for short) is
used for processing and studying on various data to get the valuable
result. For example, according to the order details, calculate and find the
goods whose sales growth rate in the recent 3 years is greater than 20%.
The data source of DCAS is
usually the structured data, such as, database, txt file, and spreadsheet. The
calculation methods include filtering, grouping, summarizing, sorting,
comparison, and discovering the correlation. Similar to ERP, CRM, Reporting
tools, Dashboard, OLAP, and ETL, DCAS is also a type of BI.
Desktop BI refers to
the BI tools running on the desktop environment, almost without any supports of
server. They usually only provides the core BI functions and requires less
dependencies on the technical environments. There is an interesting phenomenon: most
DCAS tools belong to the Desktop BI, including Excel which holds the largest
market shares in the sector of commercial BI tools, R project which ranks the
first in the open source software market. Similar examples also include StataCorp Stata, Raqsoft ES series, IBM SPSS, and MathWorks
MATLAB, etc.
Is this a coincidence? Compare their features and
you can clearly understand the root cause of this phenomenon.
If you ever read the article
of What
Role Desktop BI Plays, you should know that Desktop BI is characterized
by the below features:
l Lightweight BI tools: Desktop BI neither explores much about the
business details directly nor provides a great number of modules to give the
ready-to-use answer. Usually, a work process is required to solve a problem.
l Quicker problem
solving: Focusing on BI, the Desktop BI does not require the
technical assistance and is ideal for solving the complex problems quickly.
l Most Desktop BI users
are business-oriented, such as the accountants, banking account manager,
business analyst, and stock analyst.
l Self-service and
Independence: Desktop BI is usually used by users to complete the BI
task independently.
l Low hardware
requirement: Desktop BI is a desktop application with low hardware
requirements.
Then, let’s check the features of DCAS:
To address the temporary needs
DCAS is usually used to address the temporary needs, such as
the RStudio or esProc computation: For those clients accounting for top
50% of the total sales last year, whose ranks increased this year? The clients’ sales are usually already stored in the
business systems and may have been ranked, because these data is frequently
used. But for the data not for daily use and only be used in
specific occasions, such as “clients accounting for top 50% of the total sales”
and “year-on-year comparison based on rankings”, they are usually not
available.
The data to be frequently-used can usually be predicated in the
early stage of BI system development. The ready-to-use module can be built with
Solution BI tools such as the Report Tools, Dashboard, and OLAP. For example,
the Dashboard of QlikView is quite fit for the above-mentioned client sales
ranking or even the sales ranking.
For the data that is seldom used, since it is usually hard to
predicate and less possible to use, the cost is quite high to build all means
to get these data into the ready-to-use modules. Therefore, we need to conduct
the temporary computation. The Desktop BI refers to the lightweight tools that
do not explore into the business details. Although Desktop BI tools do not
provide the ready-to-use module to get the answer, they can be used to address
these temporary needs via calculation easily. It can be seen clearly that DCAS
is characterized by these features of Desktop BI.
To meet the sudden demands
DCAS is often used to address the sudden needs. For example,
find the product whose sales values are rising in 5 consecutive weeks through
rapid calculation in Excel or esCalc, so as to launch the marketing campaign
aimfully. Such needs are pressing since the correct results must be calculated
out in limited time. In order to achieve the goal of rapid calculation, DCAS
shall allow for the full control by users, especially the Business experts must
be capable to act independently, and the DCAS functions must focus on the BI
sectors. These are just the Desktop BI features.
On the contrary, Solution BI like SAS or SAP usually requires
the collaboration between business personnel, DBA, SQL composer, Web
administer, programmer, report script developer, and experts in several areas.
In addition, they also need going through a series of work processes like the
requirement management, departmental approval, resource provision, developing,
and responding. The timeline is completely not guaranteed at all, and thus it
is not fit for addressing the sudden needs.
What-if method
It is always easier to solve the BI problem with clear
computational goal. However, the complex problems are always abstract and
ambiguous. To address them, DCAS requires the what-if analysis method. For
example, you can resort to RStudio to find the reason for the current climbing
complaint rate. To solve such ambiguous problem, we need some reasonable
assumptions. For example, the new product debut gives rise to the laggard after-sales,
product quality drawback, and after-sale platform failure. These assumptions
are the decomposition of goal, that is, decomposing the ambiguous and great
target into several simple and clear small goals. Through validating and
calculating several simple and clear goals, the complex, ambiguous and great
goal can be solved.
The learned and experienced business expert is the key to
what-if analysis in determining:
what factors are related to the goal? Of these factors, which factors cover all possibilities
and do not overlap mutually? Which
factors can be verified explicitly? Which
factors can be further divided? What are
the weights of these factors? Which are
highly possible and which are relatively easier to verify? To make the correct judgment on these questions, you
may need the in-depth business understanding. Therefore, DCAS tools must be
business-oriented, such as esProc. Being business-oriented is just a feature of
Desktop BI.
Individual creative work
The labor can be divided into two types of the repetitive
work and the creative work. In BI sector, the repetitive work refers to those
problems that can be solved through teamwork or collaboration between multiple
persons, for example, the commonly-used reports in enterprise, OLAP model tailored
for specific industry, and classic correlation analysis. They are in the scope
of Solution BI conventionally. But the creative work is quite another thing.
For example, use RStudio or esProc to find the new product with the greatest
market potential.
For the creative work, no standardized and existing solution.
The creative work requires the rich expertise of business experts and
computation, and DCAS is really good at such computation. Different experts may
see from different perspectives and be in different positions, take different
analytical methods, and reach different conclusions. Therefore, their
respective process cannot be reproduced. Such calculation is soaked with the
strong personal style, being related to the individual background, work experiences,
and business preferences of business experts. It is the typical creative work
by individuals. The collaboration will backfire and hinder the user creativity.
Therefore, DCAS is usually adopted by users independently as
a type of typical Desktop BI.
Ability of expressing the business
Ability of expressing the business is the ability to convert
the business jargons into the computer languages. Unlike other BI tools, DCAS
users are usually required to analyze the complex goal, which demands the creative
work on the basis of a strong business background. In view of this, we can
conclude that the core ability of DCAS is to express the business ideas and
plans efficiently and cost-effectively, which is an important criterion to
discriminate the good DCAS tools from the bad ones. This core ability includes
providing the friendly UI, the business-oriented syntax rule, the intuitive and
easy-to-understand formulas, and the free analytical style. For example, with
esCalc, merge the basic salary, performance, attendance, and multiple
spreadsheets into a practical salary sheet according to the No. of employee.
So, we can say that performance is not the top priority for
DCAS. The core features of Solution BI like multi-core parallel computing,
cloud, and cluster computing can boost the performance only, but not the
ability of expressing the business. These features may backfire, distracting
users from reaching the business result and even bringing about a bad impact on
the correctness of computational results.
In addition, the normal PC can offer the more than enough
computational capability. Even the CPU released 5 years ago - Intel i7 - can
support more than 8GB memory and are still powerful enough for running almost
all DCAS. In
fact, not
having to rely on servers, most computation and analysis problems can be solved on PCs
that are believed to provide only the relatively low performance nowadays. The vast majority of DCAS belongs to the
Desktop BI. In case any computational problems requiring a higher PC performance
are encountered, DCAS tools, for example R and esProc, can also handle them
well with its advanced features, though it seldom happens.
No comments:
Post a Comment