Types of reports
D ATA C O M M U N I C AT I O N C O N C E P T S
Hadrien Lacroix
Curriculum Manager
Presentation strategy
DATA COMMUNICATION CONCEPTS
Chapter 3
How to structure a written report?
Types of reports
Reproducibility
Write precise and clear reports
DATA COMMUNICATION CONCEPTS
Written reports
Explain data analysis project
Sentiment analysis on product reviews
DATA COMMUNICATION CONCEPTS
Written reports
Explain data analysis project
Sentiment analysis on product reviews
Communicate findings
30% negative ratings for delayed shipping
Predict ratings with 90% accuracy
Standards
DATA COMMUNICATION CONCEPTS
Written reports
Explain data analysis project
Sentiment analysis on product reviews
Communicate findings
30% negative ratings for delayed shipping
Predict ratings with 90% accuracy
Standards
Give recommendations to drive change
DATA COMMUNICATION CONCEPTS
Types of reports
Informational Analytical
Factual information Analysis (relationships/recommendations)
Short Varies (short or long)
Not strict structure Strict structure
Inform about facts Data-driven decisions
DATA COMMUNICATION CONCEPTS
Final report
Elements
Data analysis
Findings and results
Visuals
Format
Long
Audience
Details
DATA COMMUNICATION CONCEPTS
Summary report
Elements
Key findings and recommendations
Visuals
Format
Short (< 5 pages)
Summary of final report
Link to main document
Audience
No need for details
DATA COMMUNICATION CONCEPTS
Report structure
Introduction
Purpose
Analysis of the product reviews gathered from website
Rating prediction based on review
Contextual information
Increase in negative reviews
Question of analysis
Factor affecting bad user experience
DATA COMMUNICATION CONCEPTS
Report structure
Introduction
Body
Data
Description and tables
Methods
NLP and Random Forest
Analysis
Visuals
Graphs with most common words
Results
Description and visuals
30% negative ratings associated with words "delayed" and "shipping".
DATA COMMUNICATION CONCEPTS
Report structure
Introduction
Body
Data
Methods
Analysis
Results
Conclusions
Restate question
Summarize important results
Add recommendations
DATA COMMUNICATION CONCEPTS
Report structure
Business context
1-3-25
1 page of abstract
≤ 3 pages of executive summary
≤ 25 pages of detail
DATA COMMUNICATION CONCEPTS
Audience
People with little time
Introduction
Conclusion
Scan body
DATA COMMUNICATION CONCEPTS
Audience
Customer or internal collaborator
Executive team
Scan introduction and conclusions
Recommendations
DATA COMMUNICATION CONCEPTS
Audience
Customer or internal collaborator
Executive team
Technical stakeholder
Body
DATA COMMUNICATION CONCEPTS
Let's practice!
D ATA C O M M U N I C AT I O N C O N C E P T S
Reproducibility and
references
D ATA C O M M U N I C AT I O N C O N C E P T S
Hadrien Lacroix
Curriculum Manager
Written report
A report must be clear and reproducible.
DATA COMMUNICATION CONCEPTS
Reproducibility example
Baking a cake
Recipe
Raw ingredients
Our oven and kitchen measuring gadgets
Cake with a similar flavor
Data project
Run analysis again - same results
DATA COMMUNICATION CONCEPTS
Replicability example
Baking a cake
Own utensils
Own ingredients
Data project
Different environment
DATA COMMUNICATION CONCEPTS
Reproducibility and replicability virtues
Prevents duplication of effort
Build upon pre-existing work
Focus on new challenges
Peer review
Tool agnostic
DATA COMMUNICATION CONCEPTS
Best practices
1. Keep track of how results were produced
Well document scripts
Comments in code
List packages and environment used
Version control
DATA COMMUNICATION CONCEPTS
Best practices
1. Keep track of how results were produced
2. Avoid manual data manipulation
Data versioning
Store raw data and intermediate steps
Adapt and resolve problems
Example: data imputation
impute missing values with the mean
save and close editor
how to know which values were replaced in the first place?
DATA COMMUNICATION CONCEPTS
Best practices
1. Keep track of how results were produced
2. Avoid manual data manipulation
3. Control randomness
Random seeds for ML pipelines
Controls confounding variables
DATA COMMUNICATION CONCEPTS
Best practices
1. Keep track of how results were produced
2. Avoid manual data manipulation
3. Document randomness
4. Interpretability
Understand the cause of a decision or predict model results
Story with compelling narrative
Link with reproducibility
1 Molnar C. Interpretable Machine Learning. 2019.
DATA COMMUNICATION CONCEPTS
Best practices
1. Keep track of how results were produced
2. Avoid manual data manipulation
3. Document randomness
4. Interpretability
5. Cite bibliography correctly
DATA COMMUNICATION CONCEPTS
References
A citation is the basic information required to identify and locate a specific publication
DATA COMMUNICATION CONCEPTS
References
Different styles but same underlying logic
Book: Author Name (Year). Title. Publisher.
Journal Article: Author Name. (Year) 'Article Title.' Journal Title, Volume Number, Issue
Number, Page Numbers.
Website: Author Name. Date of Publication, 'Title of Page/Work.' Title of Website, Location
APA style:
In text citations (author, date)
DATA COMMUNICATION CONCEPTS
Reference
Reference management tools
Easier to keep track
Change between styles
Search for reference online
Options:
EndNote
Mendeley
RefWorks
DATA COMMUNICATION CONCEPTS
References
Business context
Less strict
Simpler (hyperlink)
==> information available and retrievable
DATA COMMUNICATION CONCEPTS
Let's practice!
D ATA C O M M U N I C AT I O N C O N C E P T S
Write precise and
clear reports
D ATA C O M M U N I C AT I O N C O N C E P T S
Hadrien Lacroix
Curriculum Manager
Written report
A report must be clear and reproducible.
DATA COMMUNICATION CONCEPTS
Write precise and clear reports
Concise
Precise
Avoid misleading and confusion
Meaningful message
1 Nolan D, Stoudt S. Communicating with Data. OUP Oxford. 2021.
DATA COMMUNICATION CONCEPTS
Empty phrases
Contain no information
It is interesting to note that
The fact that
It should be pointed out that
It is well known that
It is obvious that
1 Nolan D, Stoudt S. Communicating with Data. OUP Oxford. 2021.
DATA COMMUNICATION CONCEPTS
Empty phrases
Contain no information
Distracting
==> should be removed
1 Nolan D, Stoudt S. Communicating with Data. OUP Oxford. 2021.
DATA COMMUNICATION CONCEPTS
Empty phrases
Negative ratings are associated with the Another important point is the fact that
words "delayed" and "shipping" negative ratings were associated with the
words "delayed" and "shipping"
1 Nolan D, Stoudt S. Communicating with Data. OUP Oxford. 2021.
DATA COMMUNICATION CONCEPTS
Concrete nouns
Write concrete nouns
Avoid "this", "that", "it"
Adds cognitive load
Distracts them from insights
1 Nolan D, Stoudt S. Communicating with Data. OUP Oxford. 2021.
DATA COMMUNICATION CONCEPTS
Concrete nouns
This shows an accuracy of 80% when The model shows an accuracy of 80% when
predicting customer churn. predicting customer churn.
1 Nolan D, Stoudt S. Communicating with Data. OUP Oxford. 2021.
DATA COMMUNICATION CONCEPTS
More pronouns
Active voice: emphasis on the author
Passive voice: stuffy and hard to read
Academic vs business context
1 Nolan D, Stoudt S. Communicating with Data. OUP Oxford. 2021.
DATA COMMUNICATION CONCEPTS
Redundant adjectives and adverbs
Phrases that say the same thing twice
Introduce a new
Done previously
Eliminate redundant adjective and adverbs
1 Nolan D, Stoudt S. Communicating with Data. OUP Oxford. 2021.
DATA COMMUNICATION CONCEPTS
Run-on sentences
Two or more independent clauses connected incorrectly
There is a correlation between delayed shipping and customer rating, the shipping delay
is the cause for negative review.
Correction
Make two sentences
Use dependent clause
DATA COMMUNICATION CONCEPTS
Let's practice!
D ATA C O M M U N I C AT I O N C O N C E P T S
Case study: report
on credit risk
D ATA C O M M U N I C AT I O N C O N C E P T S
Hadrien Lacroix
Curriculum Manager
Credit risk
Credit risk: probability of defaulting
Loanme bank wants to predict if a customer is likely to default
Raw data available
Data Exploration Analysis
Model training and evaluation
DATA COMMUNICATION CONCEPTS
Audience
Non-technical stakeholders
Bank decision-makers
DATA COMMUNICATION CONCEPTS
Story
Background:
Increase in defaulting percentage over last 5 years.
Predicting which customers had a high probability of default.
Insight: People with more unemployment periods tends to default more
Insight: People with lower income tend to default more
Climax: Possible to predict which people is more likely to default with an accuracy of 95%
Next steps: Run a trial on a control population
DATA COMMUNICATION CONCEPTS
Tech or non-tech
Translate technical results
DATA COMMUNICATION CONCEPTS
The right data
Audience persona
Role: Financing Department Director
Interest: Decision on implementing an automated loan rejection system
Appropriate data:
Relationship between age or income and loan default
Percentage customer defaulting over the next months
DATA COMMUNICATION CONCEPTS
Statistics
Median age and income
Percentage of change
DATA COMMUNICATION CONCEPTS
Visuals
Boxplot with age vs. default condition
DATA COMMUNICATION CONCEPTS
Visuals
Boxplot with age vs. default condition
Lineplot with % change defaulting customers
DATA COMMUNICATION CONCEPTS
Correct format
Who? Financial Department director
Why? Important decisions ahead
Content: Key findings and recommendations
Channel: Send the results before the meeting
DATA COMMUNICATION CONCEPTS
Report
Written report
Summary report or final report?
DATA COMMUNICATION CONCEPTS
Report
Summary report
Informational report vs. analytical report?
DATA COMMUNICATION CONCEPTS
Report
Summary report
Analytical report
DATA COMMUNICATION CONCEPTS
Summary report structure
Introduction
Purpose
Contextual information
Question of analysis
Body
Data
Results: Key findings
Conclusions
Restate question
Central insight
Add recommendations
DATA COMMUNICATION CONCEPTS
Let's practice!
D ATA C O M M U N I C AT I O N C O N C E P T S