Ten Habits of Great Data Analysts
This site is the companion. The full book, with the case studies and the complete set of habits, is on Leanpub as PDF, iPad, and Kindle.
The bookAbout the book
This companion supports Dr. Danielle Boyce's book Ten Habits of Great Data Analysts (Leanpub & Amazon KDP), a practical, human-centered guide to how strong analysts think, collaborate, and make decisions. The tools below are meant to be revisited and reused after each chapter.
Ten Habits of Great Data Analysts is a practical, human-centered book about how strong analysts think, collaborate, and make decisions.
Rather than focusing on a single software package or coding language, the book emphasizes durable habits that remain useful across tools, platforms, and career stages.
Core message
Tools change. Habits are forever.
What readers can expect
The book blends:
- practical habits for analytical work
- stories from real research settings
- lessons from clinical and observational data
- guidance for analysts, collaborators, and learners
- a toolkit readers can immediately put to use
Why this companion site exists
Some parts of the book are most useful when they live online too.
This companion site makes it easier to revisit and share:
- the project request form
- exploratory data analysis materials
- domain-based thinking tools
- curated resources
- chapter companion pages
OverviewThe ten habits
The site mirrors the ten-habit structure from the book so you can move between the print edition and the online companion easily.
- Understand the domain and the people
- Start with a clear analytical plan
- Explore the data before modeling
- Prioritize data quality
- Protect data security and privacy
- Communicate clearly and document your work
- Organize and structure complex data
- Respect qualitative methods and unstructured data
- Know when to do it the old-fashioned way
- Keep learning and contribute to the community
Each habit pairs with a reusable tool below, a request form, a checklist, a set of framing questions.
ToolkitData analysis request form
Use this form to start a project, clarify timelines, and build a strong analysis plan before any code is written.
ToolkitExploratory data analysis checklist
Before formal modeling begins, analysts should take time to conduct an initial exploratory review of the data set. The goal is not yet to answer the research question, but to understand how the data behave.
Review the structure of the data set
- How many observations and variables are present?
- Do variable names match the documentation or data dictionary?
- Are variables stored in the expected data types?
Examine missing data
- What percentage of values are missing for each variable?
- Are there patterns in missingness across sites, time periods, or visit types?
- Are key variables complete enough for the planned analysis?
Check value ranges
- Do numeric variables fall within plausible ranges?
- Are there impossible or highly unlikely values?
- Are measurement units consistent?
Inspect categorical variables
- Are categories coded consistently?
- Do unexpected values appear?
- Are there very small categories that may affect analysis?
Look for duplicated or inconsistent records
- Are participant identifiers unique?
- Do multiple records appear where only one should exist?
- Are dates or timestamps consistent across related variables?
Explore distributions and relationships
- What do histograms or box plots reveal?
- Do scatterplots or cross-tabulations suggest relationships?
Compare the data with the study documentation
- Does the data set match the protocol or data dictionary?
- Are there variables that appear unexpectedly or are missing entirely?
Document observations
- Record unusual patterns, inconsistencies, or quality concerns.
- Note questions that should be discussed with investigators or data managers.
ToolkitDetective questions for undocumented data sets
When documentation is incomplete, analysts can approach the data like detectives.
Ask how the data might have been created
- Was the data set built from surveys, clinical visits, or EHR data?
- Were participants followed over time?
Look for study phases or waves
- Do variable names suggest baseline or follow up measurements?
- Are repeated measurements present across time?
Identify domains
- Do clusters of variables relate to imaging, genetics, medications, or other clinical areas?
Find likely stratification variables
- Are there site, sex, mutation type, or age-group variables?
- Are these complete and consistently coded?
Look for naming clues
- Do prefixes or suffixes suggest questionnaires, modules, or scoring systems?
Inspect time patterns
- Do some variables appear only after certain years?
- Could that reflect a change in study design?
Watch for repeated records
- Are multiple rows linked to the same participant across visits?
Check whether data types make sense
- Are numeric variables stored as text?
- Do categorical variables contain inconsistent values?
ToolkitThinking in data domains
A practical way to understand large data sets is to organize variables into logical domains.
A domain is a group of variables that belong together because they describe the same type of information or originate from the same part of the data collection process.
Common domain patterns
Survey waves
- baseline assessments
- follow up surveys
- annual visits
- sub-study questionnaires
Clinical domains
- imaging or radiology
- genetics
- laboratory measurements
- pulmonary or cardiac assessments
- medication and treatment history
Participant characteristics
- age
- sex
- mutation type
- enrollment site
- disease severity
Variable types
- date variables
- continuous measurements
- categorical variables
- text fields
Why this is important
Thinking in domains helps analysts:
- reconstruct study design
- plan exploratory analysis
- organize cleaning work
- communicate more clearly with collaborators
- turn a huge variable list into a meaningful system
For readers & educatorsTeaching and discussion ideas
This book works well in classrooms, workshops, journal club style discussions, and mentoring conversations.
Good discussion questions
- Which habit feels most under-taught in formal training?
- Which habit is hardest to practice under deadline pressure?
- Where have you seen data quality or documentation problems change the direction of a project?
- How should analysts balance technical skill with judgment, communication, and ethics?
Course or workshop uses
You could assign one chapter per session and ask learners to do one of the following:
- bring a real example from their own work
- adapt one toolkit page to a current project
- identify one risk the habit helps prevent
- describe how the chapter changes what they would do next
This site can support a light-touch or structured reading experience.
A simple format for a four-session series
- Chapters 1 to 3: context, planning, and exploration
- Chapters 4 to 6: quality, privacy, communication, and documentation
- Chapters 7 to 8: complexity, qualitative methods, and unstructured data
- Chapters 9 to 10: judgment, growth, and contribution
Helpful site pages for group use
- Overview of the Habits for orientation
- Data Analysis Request Form for applied exercises
- Exploratory Data Analysis Checklist for practice
- Key Resources for optional deeper reading
ReferenceKey resources
This page gathers useful resources for literature discovery, reproducibility, real world evidence, data harmonization, and responsible clinical data science.
Literature and discovery
Reproducibility and data management
Real world evidence and data quality
OMOP and harmonization
Data privacy and metadata
Tables, figures, and reporting style
AboutAbout the author
Dr. Danielle Boyce is Principal Investigator, Real World Evidence at the ALS Therapy Development Institute. She is an experienced data engineer and analyst with graduate degrees in public health and public administration. She also serves as a faculty member at Johns Hopkins University, Biomedical Informatics and Data Science Section and holds affiliations with the University of Calgary and Emory University. In addition, she is a technical consultant for Data for the Common Good at the University of Chicago.
As the parent of a child with a rare disease, Dr. Boyce has served on several participant and caregiver advisory panels for the Patient-Centered Outcomes Research Institute, the U.S. Food and Drug Administration, as well as academic institutions, pharmaceutical companies, and nonprofit organizations. Her work focuses on helping researchers, clinicians, and communities use real world data to better understand disease and accelerate research.
She lives in upstate New York with her husband, Jim, and their four children. In her free time, she enjoys volunteering in her community, visiting museums, baking, and doing arts and crafts with her children.
Get the bookBuy the book
Ten Habits of Great Data Analysts is available through Amazon KDP and Leanpub.