Lecture 1 - Business Analytics Trends Flashcards
5V’s of Big Data
- Volume
- Variety
- Velocity
- Value
- Veracity
What is the technical challenge of data? (2)
- How to deal with volume of the data
2. How to deal with real time data processing
What is the managerial challenge with data?
To manage the data that comes in from everywhere and integrate these data.
Degenerative tendency?
To focus inward on coasts and efforts, rather than outward on opportunities, changes and threats.
Pro of big data (comes from external sources) is that it force companies to use external data instead of only internal data.
What is the most valuable thing you can do with big data?
Developing new products and services
Caveat 1: Big data and thick data
Beware of quant bias or quant addiction; don’t trust too much only on quantitative data, also look at what is not measurable.
Thick Data is?
Data in which emotions are visible and the context where the data is collected is taken into account.
- Big data needs to be supported with Thick data.
- Context can relativise or give meaning to the data.
Caveat 2: Relate to business challenges
The challenge is not to just start exploring the data but to be aware of your business goals and focus your data analytic efforts on solving these
problems and improving these decisions.
Why Big Data now? combination of different causes (4)
- Availability of massive amounts of digital data
- Combination of technical developments and societal needs
- A philosophical view: Rationalism vs empiricism
- The discovery, in science itself, of the power of data
Why Big Data now: Technological developments (6)
Radical changes in: (now vs. how it was)
1. The way elementary data are captured
o Sensors (automated) vs keyboard (human)
2. The way data is stored → business intelligence (data warehouses, data analytics etc.)
o Main memory and cloud vs hard disk
3. The way data is analyzed
o Data-driven methods vs sampling
4. The way data is provided to users
o Data logistics (keeping data together) vs data integration
5. The way data is presented → visualization
o Graphical interactive visualizations vs management reports
6. The way knowledge (business rules, models) is created → by means of machine learning/
data mining
o Learning/mining vs (labor-intensive) knowledge acquisition
Rationalism vs empiricism
Rationalism: theory, thinking and abstract knowledge
Empiricism: rely on data, empirical data
Now we are in the empirical time.
Researchers view
More data is more important than better algorithms.
The more data you use for the training of the algorithm, the better the performance of the algorithm.
How did we go from Big data to AI? (3)
- The gap between the data and what you can do with it becomes bigger.
- Data became the problem, but the problem empowered the solution.
- Big data enabled AI machine learning techniques, which become a solution of solving these gaps.
How are AI technologies driven by data and analytics? (3)
- We use historical data for training
- New data for re-training
- With the models, make predictions
Three types of learning:
- Supervised: try to predict the variable you want to learn
- Unsupervised: don’t have any initial variable to learn
- Reinforcement: learn by trying out