Data Science in Practice
Welcome
How to use this book
Learning goals
A decision-focused perspective
Recurring case: North Valley Recreation and Wellness
1
Applied Data Science Projects
1.1
What makes an applied project different?
1.2
Projects as decision systems
1.3
The applied project cycle
1.4
Integrated project evidence
1.5
From concern to analytical question
1.6
Defining success
1.7
Chapter Review
1.7.1
Chapter summary
1.7.2
Common mistakes
1.7.3
Exercises
1.7.4
Questions for discussion
2
From Organizational Concern to Analytical Question
2.1
Why questioning is analytical work
2.2
Four kinds of project questions
2.2.1
Management questions
2.2.2
Research questions
2.2.3
Analytical questions
2.2.4
Implementation questions
2.3
A questioning sequence
2.4
Question maps
2.5
Questions that uncover assumptions
2.6
Chapter Review
2.6.1
Chapter summary
2.6.2
Common mistakes
2.6.3
Exercises
2.6.4
Questions for discussion
3
Planning the Project
3.1
Why project plans fail
3.2
Plan backward from the decision
3.3
Decomposing the work
3.4
Estimating effort and calendar time
3.5
A deliverable-based timeline
3.6
Milestones and review gates
3.7
Risk management
3.8
Chapter Review
3.8.1
Chapter summary
3.8.2
Common mistakes
3.8.3
Project planning checklist
3.8.4
Exercises
3.8.5
Questions for discussion
4
Team Charters and Collaboration
4.1
What a team charter is
4.2
The components of a strong charter
4.2.1
Purpose and goals
4.2.2
Roles and responsibilities
4.2.3
Communication and meetings
4.2.4
Decisions and review
4.2.5
Files and handoffs
4.2.6
Conflict and revision
4.3
Roles without silos
4.4
Decision rights
4.5
Meetings that produce evidence
4.6
Conflict and missed commitments
4.7
Chapter Review
4.7.1
Chapter summary
4.7.2
Common mistakes
4.7.3
Charter quality checklist
4.7.4
Exercises
4.7.5
Questions for discussion
5
Interim Project Reviews
5.1
Why interim reviews matter
5.2
What an interim review should establish
5.2.1
What problem is the project addressing?
5.2.2
What has been established so far?
5.2.3
What remains uncertain?
5.2.4
Is the project on track?
5.2.5
What is needed next?
5.3
Evidence of progress
5.4
Structure of an interim presentation
5.5
Assessing trajectory
5.6
Receiving and using feedback
5.7
Chapter Review
5.7.1
Chapter summary
5.7.2
Common mistakes
5.7.3
Interim review checklist
5.7.4
Exercises
5.7.5
Questions for discussion
6
Data Responsibility and Confidentiality
6.1
Data access creates obligations
6.2
Privacy, confidentiality, security, and ethics
6.3
Minimum necessary use
6.4
De-identification is not anonymity
6.5
Generative AI and external tools
6.6
Responsible analytical use
6.7
The data lifecycle
6.8
Chapter Review
6.8.1
Chapter summary
6.8.2
Common mistakes
6.8.3
Data responsibility checklist
6.8.4
Exercises
6.8.5
Questions for discussion
7
Data Dictionaries and Codebooks
7.1
A dictionary controls meaning
7.2
Begin with the unit of analysis
7.3
Dataset-level metadata
7.4
Variable-level fields
7.5
Source, standardized, and derived variables
7.6
Missing values carry information
7.7
Building the dictionary
7.8
Using the dictionary during analysis
7.9
Chapter Review
7.9.1
Chapter summary
7.9.2
Common mistakes
7.9.3
Dictionary review checklist
7.9.4
Exercises
7.9.5
Questions for discussion
8
A Reproducible Exploratory Data Analysis
8.1
EDA is an investigation
8.2
A reproducible EDA workflow
8.2.1
Stage 1: Orient to the data
8.2.2
Stage 2: Assess data quality
8.2.3
Stage 3: Examine variables individually
8.2.4
Stage 4: Examine relationships
8.2.5
Stage 5: Add relevant dimensions
8.2.6
Stage 6: Convert findings into decisions
8.3
Missing-data decisions
8.4
Outlier decisions
8.5
Recurring case: connecting EDA to the problem
8.6
A professional EDA report
8.7
Chapter Review
8.7.1
Chapter summary
8.7.2
Common mistakes
8.7.3
Reproducibility checklist
8.7.4
Exercises
8.7.5
Questions for discussion
9
Researching Secondary Data
9.1
Why secondary evidence matters
9.2
Begin with an information need
9.3
A structured search process
9.4
Evaluating a source
9.4.1
Authority and purpose
9.4.2
Coverage and selection
9.4.3
Definition and method
9.4.4
Timeliness and revision
9.4.5
Granularity
9.4.6
Permission and citation
9.5
The source log
9.6
Combining sources
9.7
Census and geographic data
9.8
Secondary evidence as context, not substitution
9.9
Chapter Review
9.9.1
Chapter summary
9.9.2
Common mistakes
9.9.3
Source review checklist
9.9.4
Exercises
9.9.5
Questions for discussion
10
Analysis and Machine Learning
10.1
Begin with the purpose
10.2
A defensible modelling workflow
10.3
Linear regression
10.4
Logistic regression
10.5
Decision trees
10.6
Random forests
10.7
Clustering and segmentation
10.8
Evaluation must match the decision
10.8.1
Regression metrics
10.8.2
Classification metrics
10.9
R demonstration structure
10.10
Error analysis and responsible use
10.11
Chapter Review
10.11.1
Chapter summary
10.11.2
Common mistakes
10.11.3
Model quality checklist
10.11.4
Exercises
10.11.5
Questions for discussion
11
Forecasting, Projections, and Customer Value
11.1
Forecast, projection, scenario, or target?
11.2
Preparing time-series data
11.3
Moving averages
11.4
Regression forecasting
11.5
Exponential smoothing and ETS
11.6
ARIMA
11.7
Time-ordered validation
11.8
Population-utilization projections
11.9
Scenario and sensitivity analysis
11.10
Customer lifetime value
11.10.1
Introductory revenue CLV
11.10.2
Margin-based CLV
11.10.3
Retention and churn
11.10.4
Discounted CLV
11.10.5
Cohorts and segments
11.10.6
CLV in Excel
11.11
Chapter Review
11.11.1
Chapter summary
11.11.2
Common mistakes
11.11.3
Forecast and value checklist
11.11.4
Exercises
11.11.5
Questions for discussion
12
Data Storytelling and Thematic Cohesion
12.1
Exploration is not the final story
12.2
Begin with the audience
12.3
Build a central message
12.4
A client-focused narrative
12.5
Storyboards
12.6
Evidence hierarchy
12.7
Editing for signal
12.8
Personas and segment stories
12.9
Ethical storytelling
12.10
Chapter Review
12.10.1
Chapter summary
12.10.2
Common mistakes
12.10.3
Story review checklist
12.10.4
Exercises
12.10.5
Questions for discussion
13
Professional Tables and Visualizations
13.1
Begin with the comparison
13.2
Tables are analytical displays
13.3
Use perceptually accurate encodings
13.4
Reduce clutter
13.5
Direct attention
13.6
Explanatory titles and annotation
13.7
Showing uncertainty
13.8
Charts requiring caution
13.9
Geospatial visualization
13.10
Accessibility
13.11
Dashboards and Power BI
13.12
Chapter Review
13.12.1
Chapter summary
13.12.2
Common mistakes
13.12.3
Visualization checklist
13.12.4
Exercises
13.12.5
Questions for discussion
14
Reports, Recommendations, and Presentations
14.1
Reports are arguments supported by evidence
14.2
Full report structure
14.3
Short decision reports
14.4
Executive summaries
14.5
The recommendation chain
14.6
Recommendations and SMART objectives
14.7
Selecting evaluation metrics
14.8
Presentation design
14.9
Explaining technical methods
14.10
Delivery and questions
14.11
From complete to convincing
14.12
Quality assurance
14.13
Professional handoff
14.14
Chapter Review
14.14.1
Chapter summary
14.14.2
Common mistakes
14.14.3
Final communication checklist
14.14.4
Exercises
14.14.5
Questions for discussion
15
Technical Refreshers
15.1
Use refreshers when the project needs them
15.2
Statistics refresher
15.2.1
Variables and measurement
15.2.2
Centre and spread
15.2.3
Counts, proportions, and rates
15.2.4
Conditional comparison
15.2.5
Confidence intervals and tests
15.3
R and RStudio refresher
15.3.1
Project organization
15.3.2
Import and inspection
15.3.3
Transformation
15.3.4
Grouped summaries
15.3.5
Reproducible outputs
15.4
Excel refresher
15.4.1
Separate inputs, calculations, and outputs
15.4.2
References
15.4.3
Tables, lookups, and pivots
15.4.4
Formula auditing
15.5
Power BI refresher
15.5.1
Grain and relationships
15.5.2
Columns and measures
15.5.3
Date tables
15.5.4
Validation
15.6
Selecting the best tool
15.7
Reproducible software demonstrations
15.8
Chapter Review
15.8.1
Chapter summary
15.8.2
Common mistakes
15.8.3
Technical review checklist
15.8.4
Exercises
15.8.5
Questions for discussion
16
Glossary
17
Project Templates and Checklists
17.1
Project brief template
17.2
Question map template
17.3
Timeline and milestone template
17.4
Risk register template
17.5
Team charter template
17.5.1
Purpose and goals
17.5.2
Roles and review
17.5.3
Operating agreements
17.6
Meeting record template
17.7
Data dictionary review checklist
17.8
EDA report outline
17.9
Secondary-source log template
17.10
Analytical plan template
17.11
Projection assumption template
17.12
CLV assumption template
17.13
Recommendation-evidence matrix
17.14
SMART implementation template
17.15
Interim review checklist
17.16
Final report revision audit
17.17
Final deliverable checklist
18
Appendix: NVRW Complete Project Synthesis
18.1
Executive summary
18.2
Project brief
18.3
Team charter and work plan
18.4
Data package and relationships
18.5
Data dictionary and quality assessment
18.6
Exploratory analysis
18.7
Secondary evidence
18.8
Retention model
18.9
Forecasting and utilization projections
18.10
Customer lifetime value
18.11
Story and recommendations
18.11.1
Recommendation 1: Pilot early-engagement outreach
18.11.2
Recommendation 2: Review capacity at the program level
18.11.3
Recommendation 3: Incorporate utilization scenarios into facility planning
18.12
Evaluation plan
18.13
Excellent final report structure
18.14
Excellent presentation storyboard
18.15
Reproducible assets
18.16
Complete-project exercise
Data Science in Practice
17.7
Data dictionary review checklist
Dataset version, date received, dimensions, source, and unit are recorded.
Every table has a grain and key.
Every source and derived variable has one dictionary row.
Definitions explain business meaning rather than repeat names.
Types, formats, units, allowed values, and examples are clear.
Missing-value codes and counts are documented, including zero missingness.
Constraints, relationships, derivations, and ownership are recorded.
Privacy or sensitivity classifications are visible.
The dictionary matches the final analytical data.