Professional Documents
Culture Documents
Big Data will transform we live, work, and perceived the world around us. Big Data will bring quick value, is/will be impactful and be a measurable ROI for organizations.
Big data, data analytics, and human insight are far better together.
Data (potential for insights) will likely be on the balance sheets of corporations as an weighed asset. Sample statistics are inferior to N=ALL.
Cited Source: Mayer-Schonberger V. & Cukier K. (2013). Big Data: A revolution that will transform how we live, work, and think. NY, NY: Houghton Mifflin Harcourt Publishing Company. http://www.amazon.com/Big-DataRevolutionTransformThink/dp/0544002695/ref=sr_1_1?s=books&ie=UTF8&qid=1384295044&sr=11&keywords=big+data+a+revolution+that+will+tran sform+how+we+live+work+and+think
Government
Law enforcement Counter terrorism Traffic flow optimization
Telecom
Broadcast monitoring Churn prevention Advertising optimization
Manufacturing
Supply chain optimization Defect tracking RFID Correlation Warranty management
Energy
Weather forecasting Natural resource exploration
Healthcare
Drug development
Scientific research Evidence based medicine Healthcare outcomes analysis
Anti-money laundering
Risk management
Particulars
500 million mathematical models applied. Relied on the concept of correlation NOT causality. Models continuously experimented. 50m common key words searches analyzed. 45 Key word search compared against CDC list of flu outbreaks from 2003-2008.
Googles model found a strong correlation between their predictions and the official figures nationwide.
Big Data was the mechanism and likely the better tool to combat the next pandemic.
http://www.google.org/flutrends/us/#US-NV
Particulars
Transparency forces vendor to mange their own inventory Walmart increasingly does not take ownership of the product until the point of sale. Walmart uses correlations to uncover consumer buying habits
Reduce risks and cost of inventory ownership to Walmart Improve shopping experience to the trends, taste, and needs of consumer quickly Hurricane preparations- Flashlights, pop-tarts (strawberry #1), sugar-breakfast snacks are the top sellers together.
Summary
Particulars
Individual genome sequencing approached $1000.00 in the US. Usually, a single specific marker (weakness) is evaluated within a sample of a persons generic code (small portion). New marker, new sample of DNA, another $1000.00 Iconic CEO of Apple Diagnosed with pancreatic cancer in 2004 Liver transplant in 2009 One of the first persons to have his entire DNA sequenced and of his cancer tumor Entire genetic code available to his doctors to specialize treatment options that was individualized for Mr. Jobs. Performing analytics of the entire genetic code of the patient, not just a sample or specific marker.
Particulars
More (All) Data is Better (N=All) Big Data is Messy | Not Exact Correlation versus Causality Datafication Value Implications
Big Data relies more and more on all the information (as reasonably feasible)
More of the dataset can reveal more detail and provide a clearer perspective typically hidden from just sampling Credit card companies are looking at anomalies within the entire transactional dataset for fraud and abuse (near real-time / real time)
Technology and techniques allow for analyzing more data than just a small sample size
DNA sequencing Google Flu Predictive models / trends Information technology advances has unleashed the power to digitizing the big data analysis Constant changing and tweaking the algorithm models to meet real world dynamics
Private entities and individuals can now have access to vast amounts of data for analysis
Democratization of data and information Ever-increasing social media churning machine driven by human nature / sentiment - twitter feeds Machine telemetric / M2M (data / log exhaust) from cellphones, web clicks, and sensor feeds
Correlation
Datasets are too big for simple cause and effect. Departure from past where experts using hypothesis driven by theories about how something works or some impending event. Quantifies the statistical relationship between two values. (if value A , then likely value B ) Correlation is not certain, only probable. Allows for us to predict to a certain level of likelihood. Now by using N=ALL (data) we can leverage a data driven analysis Less Bias More accurate Todays technology and software/algorithms make correlation with big data possible.
Primary Use
Recombinant (fusing)
Supermarket PoS Data and Social Media
Extensibility of Data
Retail Surveillance Cameras