Download presentation
Presentation is loading. Please wait.
Published byCamilla Washington Modified over 9 years ago
1
SUPPLY CHAIN OF BIG DATA
2
WHAT IS BIG DATA? A lot of data Too much data for traditional methods The 3Vs Volume Velocity Variety
3
UNIQUE PROBLEMS OF BIG DATA Data is not always human-readable Traditional storage and processing methods are too slow Applications of data change
4
THE RISE OF BIG DATA 1 petabyte(PB) = 1,000 terabytes = 1,000,000 gigabytes Average consumer hard drive is 500GB Google processes 100PB of data a day EBay processes 100PB of data day Facebook has 300PB of data, growing by 600TB per day
5
THE SUPPLY CHAIN OF BIG DATA Collecting Storing Processing Applications Analysis Machine Learning
6
COLLECTING Automated Sensors in a machine Analytics on a website Public APIs Publicly available Limited Can be free or paid Buying Data Generally unlimited DataSift, FullContact
7
STORING Relational databases Slow after a certain point “sharding” or distributed computing only does so much NoSQL Key: Value stores, Document stores Distributing the database is effective MongoDB, CouchDB, Cassandra Cloud Solutions Google Cloud Computing, AWS Cloud solutions are cheap and managed Designed to be fast and have high uptimes
8
PROCESSING Cleaning Missing Data Outliers Sampling Is big data true population? Do you need every bit of information
9
ANALYZING Visualization Tableau, Qlik Easy to learn Programming R, Python Tailored to the company Distributed Computing Hadoop, Spark, AWS
10
MACHINE LEARNING Machine learning is making predictions based on data Includes many algorithms Can be supervised or unsupervised Pitfalls Poor data makes for poor predictions
11
ARTIFICIAL NEURAL NETWORKS (ANN) Subset of machine learning Contains one or more layers
12
GOOGLE BRAIN Deep Neural Network (DNN) Analyzed 10 million images from YouTube videos Learned to identify a…
13
APPLICATIONS OF MACHINE LEARNING Financial Trading Advertising Fraud Detection Computer Vision Natural Language Processing
14
ETHICS Is collecting data ethical? “Always listening” Security How safe is everyone’s data? Who has access to the data?
15
EXAMPLES Craiglist Missed Connections Craiglist Missed Connections How Big is Snapchat? How Big is Snapchat? 18 th and 19 th century ship logs 18 th and 19 th century ship logs Neural Network Painting Neural Network Painting
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.