Computing Concepts

Leveraging computer
Focus Point tools to extend human
potential
 Computatio
”Training Wheel” ”Truly Complimentary” Competitive
nal artifacts
Cognitive Artifacts Cognitive Artifacts Cognitive Artifacts

These are artifacts that These are artifacts that These are artifacts that
compliment human amplify and improve our amplify and improve our
intelligence in such a way abilities to perform cognitive abilities to perform cognitive
that their use amplifies and tasks in a way that is truly tasks when we have use of
improves our ability to complementary, i.e., you are the artifact but when we take
perform cognitive tasks and better with the artifact, but away the artifact we are no
once the user has mastered the artifact cannot replace or better (and possibly worse) at
the physical artifact, they can compete with you. E.g., pen performing the cognitive task
use a virtual/mental and paper. than we were before.
equivalent to perform the
same cognitive task at a
similar level of skill, e.g., an
abacus.
 Computational problem solving
Making mistakes Pattern Recognition
I can enjoy things that go wrong and Is this like a problem I’ve already
learn from them. I see mistakes as a solved? How is it different? How are
normal part of solving problems. the parts of the problem connected?

Perseverance The Computational Thinker: Decomposition

I don’t give up. I’m prepared to
keep having a go to see what
Attitudes and Skills Can I break this problem up
into smaller parts? Can I
happens. I keep going, even when explain the different parts of
things seem confusing. I’m
Attitudes

this problem and solution?

determined to find solutions.

Skills
Imagination Algorithm Design
I can look at things in unusual ways. What do I need to think about
I’m ready to consider the impossible. to make this happen? What are
Sometimes I leave a problem for a the steps I will need to do to
while. A solution might come to me solve this problem?
when I’m thinking about something
else. Problem solving
Designing systems
Collaboration Understanding behavior Abstraction and Generalization
I can use other people’s ideas. I Which is the information I actually need?
can share my ideas. We can talk What don’t I need to know? Have I made
together to solve a problem. I this more complicated than I need to?
can teach my peers and they can Will this work for other things?
teach me.
 Computational thinking

Decomposition Pattern recognition Algorithms Abstraction

is the breaking down of a is the identification of are sequences of is the generalization of a
complex problem into a common characteristics instructions used to solve problem so that a solution
series of smaller or simpler among different problems problems or perform tasks can be applied to different
problems. and within the same in a structured manner. specific problems.
problem. Recognizing
these patterns can further
simplify the problem.
Decomposition

1 2 3 4 5 6

Identify the Break down Create a Assign Solve each Integrate the
problem the problem hierarchy responsibilities subtask solutions
Pattern recognition techniques and
approaches
Machine Learning: Machine learning algorithms can
Statistical Pattern Recognition: This approach involves automatically learn patterns from data and make
using statistical models and techniques to analyze and predictions or classifications based on the learned
classify patterns based on their statistical properties. It patterns. This approach often involves training a model
includes methods such as Bayesian classifiers, hidden on a labeled dataset and using it to classify new,
Markov models, and statistical feature extraction. unseen data. Examples include decision trees, support
vector machines, and neural networks.

Computer Vision: Computer vision techniques aim to Natural Language Processing: In the context of
extract meaningful patterns or structures from visual language and text analysis, pattern recognition
data, such as images or videos. This includes tasks like techniques are used to extract linguistic patterns,
object detection, image segmentation, and facial sentiment analysis, text categorization, and
recognition. information extraction.
Key Characteristics of Algorithms:

WELL-DEFINED: ALGORITHMS FINITE: ALGORITHMS MUST DETERMINISTIC: ALGORITHMS EFFECTIVE: ALGORITHMS ARE GENERALIZABLE: ALGORITHMS
HAVE A CLEAR AND HAVE A FINITE NUMBER OF ARE DETERMINISTIC, MEANING DESIGNED TO BE EFFECTIVE, CAN BE APPLIED TO MULTIPLE
UNAMBIGUOUS DESCRIPTION STEPS. THEY SHOULD THAT GIVEN THE SAME INPUT, MEANING THAT THEY SOLVE INSTANCES OF A PROBLEM.
OF THE STEPS TO BE EVENTUALLY TERMINATE AND THEY WILL ALWAYS PRODUCE THE PROBLEM OR ACHIEVE THEY CAN BE DESIGNED TO
FOLLOWED. EACH STEP PRODUCE AN OUTPUT WITHIN THE SAME OUTPUT. THERE THE TASK THEY ARE INTENDED HANDLE VARIOUS INPUTS AND
SHOULD BE PRECISELY DEFINED A FINITE AMOUNT OF TIME, SHOULD BE NO AMBIGUITY OR FOR. THEY SHOULD PROVIDE A ADAPT TO DIFFERENT
AND EXECUTABLE. EVEN IF THE INPUT IS LARGE. RANDOMNESS IN THE STEPS. CORRECT AND RELIABLE SCENARIOS.
SOLUTION.
 • Problem Abstraction:
• involves understanding a problem or task at a high level, identifying the key
objectives, and defining the requirements. It focuses on the problem domain
and the desired outcomes without considering implementation details.

• Data Abstraction:
• involves defining and working with data types, structures, and objects without
concerning ourselves with the internal representation or implementation
details. It allows us to encapsulate data and operations into abstract data
types, providing a clear interface for interacting with the data.

Abstraction • Procedural Abstraction:

• involves breaking down a complex task into smaller, more manageable
procedures or functions. It allows us to define reusable code blocks that
perform specific operations, making the code more modular, readable, and
maintainable.

• Control Abstraction:
• involves managing the flow of execution and control structures in a program. It
allows us to use control flow constructs such as loops, conditionals, and
subroutines to control the order of execution and achieve desired outcomes.
Collaboration
Tools for
collaborati
on

This Photo by Unknown author is licensed under CC BY.

 MODERN DATA SCIENTIST
Data Scientist, the sexiest job of the 21 st century, requires a mixture of multidisciplinary skills
ranging from an intersection of mathematics, statistics, computer science, communication, and
business. Finding a data scientist is hard. Finding people who understand who a data scientist is, is
equally hard. So, here’s a little cheat sheet on who the modern data scientist really is.
MATH & PROGRAMMING
STATISTICS & DATABASE
✩ Machine learning ✩ Computer science fundamentals
✩ Statistical modeling ✩ Scripting language, e.g., Python
✩ Experiment design ✩ Statistical computing package, e.g., R

Collaborati
✩ Bayesian inference ✩ Databases SQL and NoSQL
✩ Supervised learning: decision ✩ Relational algebra
trees, random forests, ✩ Parallel databases and parallel query

on in data ✩
logistical regression
Unsupervised learning:
clustering, dimensionality
✩
✩
processing
MapReduce concepts
Hadoop and Hive/Pig

science ✩
reduction
Optimization: gradient
descent and variants
✩
✩
Custom reducers
Experience with xaaS like AWS

DOMAIN COMMUNICATION
KNOWLEDGE & VISUALIZATION
✩ Able to engage with senior
& SOFT SKILLS management
✩ Story telling skills
✩ Passionate about the ✩ Translate data-driven insights into
business
decisions and actions
✩ Curious about data
✩ Visual art design
✩ Influence without authority
✩ R packages like ggplot or lattice
✩ Hacker mindset
✩ Knowledge of any of visualization tools,
✩ Problem solver This Photo by Unknown author is licensed under CC BY.
e.g., Flair, D3.js, Tableau
Abstraction in programming

Charact
er
Decimal

Binary
Computer models
and simulations
• A model is a simplified representation or
abstraction of a more complex object, system, or
phenomenon.

• A simulation is a model that changes over time.

Models and simulations as computational
artifacts

Experimentation and
Prediction and Analysis
Testing

Visualization and
Education and Training
Communication
data
Data analytics
Types of data analytics
• Descriptive data analytics seeks to describe data, providing
insight and knowledge.
• Predictive data analytics seeks to make predictions from data.
• Prescriptive data analytics seeks to make decisions
(prescriptions) based on data.
Challenges in big data

STORAGE QUALITY STRUCTURE PRIVACY SCALABILITY

CONTROL
 Data
visualization
 Uses of
data
visualizati
on

This Photo by Unknown author is licensed under CC BY-SA.

Consideratio
ns for data
visualization
 Type of data
Qualitative Quantitative

Nominal Ordinal Binary Discrete Continuous

• Variables with • Variables with • Variables with • Aka Attribute data • Continuous data is
no inherent an ordered only two • Discrete data is information that can be
order of series options. information that can be measured on a continuum
ranking • E.g., Blood • E.g., Pass/Fail, categorized into a or scale
sequence. Group, Yes/No, etc. classification. • Continuous data can have
• E.g., Gender, Performance, • Discrete data is based on almost any numeric value
Race, etc. etc. counts. and can be meaningfully
• Finite number of values is subdivided into finer and
possible and the values finer increments.
cannot be subdivided • E.g., Length, Size, width
meaningfully
Thank you!
Prof Stager