Académique Documents
Professionnel Documents
Culture Documents
Big data is set to offer companies burdened with ever-growing requests access to information in a form they can
tremendous insight. But with terabytes for data, ad hoc analyses and one- easily understand and share with others.
and petabytes of data pouring in off reports. Decision makers become
to organizations today, traditional frustrated because it takes hours or This begs the question: How do you
architectures and infrastructures are days to get answers to questions, if at present big data in a way that business
not up to the challenge. IT teams are all. More users are expecting self-service leaders can quickly understand and
use? This is not a minor consideration.
Mining millions of rows of data creates
Data visualization is becoming an increasingly a big headache for analysts tasked with
sorting and presenting data.
important component of analytics in the age
Organizations often approach the
of big data. problem in one of two ways: Build
samples so that it is easier to both
analyze and present the data, or create
template charts and graphs that can
accept certain types of information.
Both approaches miss the potential
for big data.
1 Meeting the need for speed 3 Addressing data quality a chart. Outliers typically represent about
In todays hypercompetitive business Even if you can find and analyze data 1 to 5 percent of data, but when youre
environment, companies not only have quickly and put it in the proper context working with massive amounts of data,
to find and analyze the relevant data they for the audience that will be consuming viewing 1 to 5 percent of the data is
need, they must find it quickly. Visual- the information, the value of data for rather difficult. How do you represent
ization helps organizations perform decision-making purposes will be those points without getting into plotting
analyses and make decisions much jeopardized if the data is not accurate issues? Possible solutions are to remove
more rapidly, but the challenge is going or timely. This is a challenge with any the outliers from the data (and therefore
through the sheer volumes of data and data analysis, but when considering the from the chart) or to create a separate
accessing the level of detail needed, all volumes of information involved in big chart for the outliers. You can also bin
at a high speed. The challenge only data projects, it becomes even more the results to both view the distribution of
grows as the degree of granularity pronounced. Again, data visualization data and see the outliers. While outliers
increases. One possible solution is will only prove to be a valuable tool if the may not be representative of the data,
hardware. Some vendors are using data quality is assured. To address this they may also reveal previously unseen
increased memory and powerful parallel issue, companies need to have a data and potentially valuable insights.
processing to crunch large volumes of governance or information management
data extremely quickly. Another method process in place to ensure the data is Conclusion
is putting data in-memory but using a clean. Its always best to have a pro-
As more and more businesses are
grid computing approach, where many active method to address data quality
discovering, data visualization is be-
machines are used to solve a problem. issues so problems wont arise later.
coming an increasingly important
Both approaches allow organizations to
component of analytics in the age of big
explore huge data volumes and gain
4 Displaying meaningful results data. The availability of new in-memory
business insights in near-real time.
Plotting points on a graph for analysis technology and high-performance
becomes difficult when dealing with analytics that use data visualization is
2 Understanding the data extremely large amounts of information providing a better way to analyze data
It takes a lot of understanding to get or a variety of categories of information. more quickly than ever. Visual analytics
data in the right shape so that you can For example, imagine you have 10 billion enables organizations to take raw data
use visualization as part of data analysis. rows of retail SKU data that youre trying and present it in a meaningful way that
For example, if the data comes from to compare. The user trying to view 10 generates the most value. Nevertheless,
social media content, you need to know billion plots on the screen will have a hard when used with big data, visualization
who the user is in a general sense time seeing so many data points. One is bound to lead to some challenges.
such as a customer using a particular way to resolve this is to cluster data into If youre prepared to deal with these
set of products and understand what a higher-level view where smaller groups hurdles, the opportunity for success
it is youre trying to visualize out of the of data become visible. By grouping the with a data visualization strategy
data. Without some sort of context, data together, or binning, you can more is much greater.
visualization tools are likely to be of less effectively visualize the data.
value to the user.