bds14 big data analytics to the masses
DESCRIPTION
Slides from my talk at Big Data Spain 2014 in Madrid. In this talk, we will discuss our approach to bring large scale deep analytics to the masses. R is an extremely popular numerical computer environment, but scientific data processing frequently hits its memory limits. On the other hand, system to execute data intensive tasks like Hadoop or Stratosphere are not popular among R users because writing programs using these paradigms is cumbersome. We present an innovative approach to overcome these limitations using the Stratosphere/Apache Flink big data platform by means of a R package and ready-to-use distributed algorithm. This solution allows the user, with small modifications in the R code, to easily execute distributed scenarios using popular machine learning techniques. We will cover the implementation details of the proposed solution including the architecture of the system, the functionality implemented and working examples. In addition, we will cover what are the differences between our approach and other solutions that integrate R with Hadoop or other large-scale analytics systems. Finally, the results of the performance tests show that this solution is competitive with the already existing R implementations for small amounts of data and able to scale-up to gigabyte level.TRANSCRIPT
Big Data Analytics to the masses
Why it has failed and how we can fix it
Jose Luis Lopez Pino @jllopezpino
Who am I?
BI Consultant
Large-Scale & Distributed
Founding
Data Engineer
Big Data is like Tourism But if you aren’t an expert,
you can’t make the most of itIt seems easy to do
Struggle to analyze Big Data
Harlan Harris, Sean Murphy, and Marck Vaisman. Analyzing the Analyzers: An Introspective Survey of Data Scientists and Their Work. O’Reilly Media, Inc., 2013Also: Sean Kandel, Andreas Paepcke, Joseph M Hellerstein, and Jeffrey Heer. Enterprise data analysis and visualization: An interview study. Visualization and Computer Graphics, IEEE Transactions
Tools
Volker Markl. Breaking the chains: On declarative data analysis and data independence in the big data era. Proceedings of the VLDB Endowment, 7(13), 2014
Tools (Now)
Original: Volker Markl. Breaking the chains: On declarative data analysis and data independence in the big data era. Proceedings of the VLDB Endowment, 7(13), 2014
Deep analytics
Libraries!
We need libraries...
Query languages
Write your own MR/RDD/Transformations
… comprehensive ones!
Say it with memes!
When you doDeep analytics in small data
using R and CRAN packages
When you dodeep analytics in BIG data
using R and CRAN packages
When you try to program it using MapReduce
When you try to program it using Apache Spark /
Apache Flink
When you try to use a library scalable to large data sets
Can’t we do it better?
- Make it similar to normal R programs.
- Hide complexity.- Make file manipulation easier.- Part of the computing in the
cluster and part of the computer in the client.
Our approach
Our approach
Behind the scenes: Before
Behind the scenes: After
Without writing significantly different code
Competitive or even faster than R native code in small data
Competitive even in highly iterative programs in small data
And it scales
Some relevant findings
- Transmission time was not significant.- Stratosphere/Flink was competitive even in
small datasets.- Changes in the code were required.- Ensemble scenarios are the most exciting
ones.
4 Takeaways from this talk
- We still need to bring Big Data to the right people in the right place.
- We need comprehensive libraries.- We need to move data back and forth.- Use a syntax that the users are familiar with.
That’s all!- Have you found this talk interesting?
- Follow me: @jllopezpino- Looking for a job? (SEM Data Analyst,
Senior Analyst)- GYG is hiring:
- Are you interested in Data + Energy?- Keep in touch: