TDWI 17 Munich - Are enterprises ready for the 4th industrial revolution? - Santiago Cabrera-Naranjo
- 1. 1 THINK BIG VELOCITYTM = SPEED + DIRECTION With <1% of global data being analysed. Are enterprises really ready for the 4th Industrial Revolution? Santiago Cabrera-Naranjo Principal Manager May 2017
- 2. 2 THINK BIG VELOCITYTM = SPEED + DIRECTION Case Study: Data Opportunities
- 3. 3 THINK BIG VELOCITYTM = SPEED + DIRECTION 10% Of Flight Delays Case Study: Data Opportunities $8 B In additional Costs
- 4. 4 THINK BIG VELOCITYTM = SPEED + DIRECTION
- 5. 5 THINK BIG VELOCITYTM = SPEED + DIRECTION 60 k Delays & Cancellations Prevented Case Study: Data Opportunities 7 M Passengers at Destination on-time Source: General Electric
- 6. 6 THINK BIG VELOCITYTM = SPEED + DIRECTION Last year humanity generated more data than in whole human history
- 7. 7 THINK BIG VELOCITYTM = SPEED + DIRECTION By 2020 more than 50 billion things, ranging from cranes to coffee machines, will be connected to the internet.
- 8. 8 THINK BIG VELOCITYTM = SPEED + DIRECTION This means 44 zettabytes of data will be loaded annually.
- 9. 9 THINK BIG VELOCITYTM = SPEED + DIRECTION To prepare for this onslaught, some IT leaders are urging the creation of Data Lakes
- 10. 10 THINK BIG VELOCITYTM = SPEED + DIRECTION • A central repository with trusted, consistent data • Cost Efficiencies by offloading analytical systems and archiving cold data • Derive value quickly with easier discovery and prototyping • A laboratory for experimenting with new technologies and data Goals for a Data Lake © 2016 Think Big, A Teradata Company
- 11. 11 THINK BIG VELOCITYTM = SPEED + DIRECTION 1. Data Driven Businesses need Business driven Data Projects. 2. Lack of DevOps Mentality along the whole analytical lifecycle. 3. Lack of skills and capability of organization. Where are Companies failing?
- 12. 12 THINK BIG VELOCITYTM = SPEED + DIRECTION This conversation often turns straight to Artificial Intelligence and Deep Learning What is next in Big Data?
- 13. 13 THINK BIG VELOCITYTM = SPEED + DIRECTION Today data science is all too often a process where new insights and models get developed as a one-time effort or deployed to production on an ad hoc basis, and require regular babysitting for monitoring and updating Analytics Operations Mentality
- 14. 14 THINK BIG VELOCITYTM = SPEED + DIRECTION14 Getting Closer to the 4th Industrial Revolution? To get there, companies will require cross- functional teams with the right software and discipline to enable data scientists, engineers, product managers, and domain experts to all work together to create a continuous cycle that drives value to the business.
- 15. 15 THINK BIG VELOCITYTM = SPEED + DIRECTION Why Analytics Ops Only 15 percent of big data projects successfully deploy into production “ ’’- Gartner 2015 Process Engineering Analytics Ops Three distinct components are required to realize value to business At Think Big Analytics we understand that value to business is only realized from sustainable data products running in production systems. Building for production implies embedding with existing and new business processes.
- 16. 16 THINK BIG VELOCITYTM = SPEED + DIRECTION •Road objects, traffic and accident events manually reported or not at all •Multiple systems using video and sensors in vehicle – Smart navigation and safety – Sign and price recognition – Vehicle comfort – Smart parking assist •Tools: TensorFlow, Darknet, Caffe, MXNet •Techniques: – Object Detection: You Only Look Once (YOLO) v2, MultiNet, R-CNN, Single Shot MultiBox Detector (SSD), – Scene Labeling: Convolutional Neural Network, MultiNet Large Japanese Automotive Conglomerate Real-Time Streaming Streaming Results Traffic Data Service Navigation Update Darknet/Darkflow – Object Detection TensorFlow – Scene Labeling Cloud GPU Based Training TF Serving Cloud GPU Based Inference Model Updates
- 17. 17 THINK BIG VELOCITYTM = SPEED + DIRECTION Benefits ▪ Cost savings of USD millions every month. ▪ False positives reduced by 50% and detection rate increased by 60%. ▪ Transaction latency performance of just ~20 mins across 30 million transactions annually Saving millions with Artificial Intelligence (AI) © 2017 Teradata FINANCIAL SERVICES – Leading Bank Business challenges ▪ To improve fraud detection in business transactions by using deep AI. ▪ An outdated ‘human-written’ rules engine. ▪ Low fraud detection rates and worrying fraud false positives. ▪ The bank needed a trusted partner with proven expertise in deep learning and building architectures. . Solution ▪ Built advanced analytics models and created a blueprint for real-time model scoring. ▪ Deployed advanced analytics through deep learning. ▪ Became a trusted advisor and took a leading role in AI implementation and future projects. VELOCITY ™ – Artificial Intelligence Strategy Tools & Technologies
- 18. 18 THINK BIG VELOCITYTM = SPEED + DIRECTION “Innovation is not just about technology per se. It is more about new models of social organisation...” Thank you!
Editor's Notes
- #3: To prepare for this onslaught, some IT leaders are urging the creation of “data lakes.” These are centralized repositories based on Hadoop that draw raw data from source systems and then pass them to downstream facilities for utilization by the knowledge workforce. Data lake designs vary, but they typically involve data stored in Hadoop’s distributed file system (HDFS) and accessed by YARN applications such as MapReduce, Spark, Storm, Solar, Hive, HBase etc.
- #6: Data only in the US
- #10: To prepare for this onslaught, some IT leaders are urging the creation of “data lakes.” These are centralized repositories based on Hadoop that draw raw data from source systems and then pass them to downstream facilities for utilization by the knowledge workforce. Data lake designs vary, but they typically involve data stored in Hadoop’s distributed file system (HDFS) and accessed by YARN applications such as MapReduce, Spark, Storm, Solar, Hive, HBase etc.
- #12: To prepare for this onslaught, some IT leaders are urging the creation of “data lakes.” These are centralized repositories based on Hadoop that draw raw data from source systems and then pass them to downstream facilities for utilization by the knowledge workforce. Data lake designs vary, but they typically involve data stored in Hadoop’s distributed file system (HDFS) and accessed by YARN applications such as MapReduce, Spark, Storm, Solar, Hive, HBase etc.
- #13: To prepare for this onslaught, some IT leaders are urging the creation of “data lakes.” These are centralized repositories based on Hadoop that draw raw data from source systems and then pass them to downstream facilities for utilization by the knowledge workforce. Data lake designs vary, but they typically involve data stored in Hadoop’s distributed file system (HDFS) and accessed by YARN applications such as MapReduce, Spark, Storm, Solar, Hive, HBase etc.
- #16: To understand the Why of Analytics Ops, we bring back the result of the 2015 Gartner study into Big Data products. We want to be part of that 15%, not part of the 85%.