Data Communication Concepts – Datacamp. Bring your data to life. Improve your presentation and learn how to translate technical data into actionable insights.
You’ve analyzed your data, run your model, and made your predictions. Now, it’s time to bring your data to life! Presenting findings to stakeholders so they can make data-driven decisions is an essential skill for all data scientists. In this course, you’ll learn how to use storytelling to connect with your audience and help them understand the content of your presentation—so they can make the right decisions. Through hands-on exercises, you’ll learn the advantages and disadvantages of oral and written formats. You’ll also improve how you translate technical results into compelling stories, using the correct data, visualizations, and in-person presentation techniques. Start learning and improve your data storytelling today.
1. Storytelling with Data
Let’s start with the importance of data storytelling and the elements you need to tell stories with data. You’ll learn best practices to influence how decisions are made before learning how to translate technical results into stories for non-technical stakeholders.
1.1. Fundamentals of storytelling
The story begins
You recently started working as a data scientist at a company named Communicatb. For your first project, you and your team need to analyze churn customer data for a cell phone company. The goal is to predict their behavior and help develop a program to retain customers.
Your team lead knows you are an expert on storytelling. She asks you to explain to the team why crafting a compelling story is important when delivering results. You write down a list of reasons to be prepared.
One of the statements you wrote is false. Can you select which one is it?
- It will be easier for the audience to remember an anecdote on why customers churn than the correlation coefficients between customer traits.
- Your findings will be better aligned with change-adverse stakeholder expectations. They will be most likely to implement the program to retain customers.
- Even if your data do not reveal a distinct customer behavior, storytelling might influence stakeholders to create the retention program.
- The marketing team will have a better understanding of the impact of your model. It is central since they are creating the retention program.
Correct! Data is the foundation of your story. So you need to make sure you are telling a story based on reliable and accurate results!
Building a story
You nailed it! Your explanation of why stories are efficient when conveying insights went very well! Now, your team lead would like you to give a short presentation. You’re going explain the different steps involved in telling a story with data to the team. It will be the starting point for delivering the results of the churn project when ready.
You know it is an important task. To prepare your talk, you look for your notes on the storytelling course you took, but realize that some parts are erased. So you need to remember the do’s and dont’s of data storytelling.
Which of the following statements about effective data storytelling are true, and which are false?
Correctly classify the statements as either true or false.
Congratulations! You can now identify the elements that make a good data story. Relevant and reliable data as well as a compelling narrative are central to storytelling. Keep going and learn how to translate technical results for non-technical stakeholders.
1.2. Translating technical results
A non-tech story
The exploratory data analysis on the churn project is finished! It’s now time for the monthly update meeting. You will have to explain your results to the operation specialist and the program director. You are addressing a non-technical audience, and want to make sure that your presentation is adapted to the audience you’re addressing so that your message gets across.
You write down some statements you could use to explain your work, but you believe some of them are more suitable for a non-technical story, while others are too technical to include.
Can you select which sentences you should use in this case?
Excellent Job! Communicating your stories to a non-technical audience can help inspire collaboration and solve problems together, which can benefit the data-driven decision-making process.
The meeting was a success! The program director asks you to send your results to the business specialists. You need to write a report and send it by email by the end of the week. You have never met them, so you ask for their background and goals.
After you gather data, you realize you will be communicating your results to a different audience. You want them to understand your results.
Can you select which of the following examples are best practices to translate your results?
Amazing job! You’ve now mastered how to translate technical results to convey a clear and convincing story with data! Keep going and learn how to impact the decision-making process.
1.3. Impacting the decision-making process
Is it a true story?
You have done an amazing job explaining your exploratory data analysis on the churn project. Now, it’s time to run the model to predict customer churn. You know that you will have to craft an effective story to present these results.
You want to be prepared. So you read your notes on how to build a compelling narrative. But you realized that one of your notes is not accurate.
Which of the following statement is false?
- A compelling narrative is key to presenting relevant insights to your target audience in a meaningful and impactful way.
- Because you should shape the narrative to your target audience, showing only key points or findings is a good practice.
- Unless you have a great data groundwork to support your central insight, your findings will need a well-formed and compelling narrative to drive action and change.
That’s right! A strong compelling narrative is key when telling stories with data. It can make a difference in the decision-making process. But remember that the story should always be based on strong and trustworthy results.
Structured to impact
Your project on customer churn is done. You analyzed the data and built your model. You followed the steps for storytelling. Now, it’s time to structure your story to have an impact at the decision-making level. You want stakeholders to follow your recommendations.
You like to write things down. So you take a pen and paper, and write down the different things you want to say in order on sticky notes. The window suddenly opens, throwing all of your notes on the floor.
Can you organize the steps for telling a story with data that is solid enough to influence the decision-makers?
Order the steps chronologically: the first step should be on top and the last step at the bottom.
Good job! You have organized a strong compelling narrative for your story! This makes it easier to understand and can help to persuade stakeholders to take actions and drive change!
A story to compare
Great job organizing your narrative structure! The next step is to think about how you will present your insights.
You start reading and discover there are several ways to present data stories. You can compare your data, show correlation, cluster your data…
You are curious to know what type of data story would be a good fit for your data. You write down the central finding, your insights and the supporting evidence.
Can you classify your findings into the following categories?
Excellent job! You have reached the end of Chapter 1. You’ve now mastered the basics of storytelling. In the next chapters, you will build on this foundation to prepare the data to write a report or build an oral presentation
2. Preparing to communicate the data
Deepen your storytelling knowledge. Learn how to avoid common mistakes when telling stories with data by tailoring your presentations to your audience. Then learn best practices for including visualizations and choosing between oral or written formats to make sure your presentations pack a punch!
2.1. Selecting the right data
The truth about salaries
Your predictive model for customer churn, which you worked on in Chapter 1, has been deployed. Your project manager asks you to work on a new internal project. The goal is to analyze a database with employee salaries in San Francisco, USA.
After doing an exhaustive exploratory data analysis, you have to present your findings to the human resources team. They want to compare San Francisco salary growth to the one at the company; they need to understand how to forecast salaries for the next year. You are about to copy the graphs from your analysis. Your manager reminds you to select the right data for your stakeholders.
You start by writing down what you believe can help you choose the proper findings.
One of the statements you wrote is false. Can you select which one it is?
- The human resource team would likely be interested in knowing how the average salary has been increasing in the last 10 years in San Francisco.
- The human resource team has no knowledge of data analysis techniques, so code shouldn’t be included when listing the top 5 job titles.
- Select categorical data, such as the salaries on the top 10 rated companies in industry the company evolves in, that provides context to support the idea of the increased salaries.
- Select all collected numerical data about San Francisco salaries and show them in a big dashboard so it helps understand in detail why salaries have been increasing.
That’s right! You should include only the data that is relevant for your specific audience. Too much information can be overwhelming and can cloud the understanding of the salary increase rate.
Well done! Your presentation with the human resource department was a success. Your team lead asks you to show your data analysis results to different stakeholders. Before you dive into preparing the presentation or the report, you want to make sure that you are aligned with their interests.
With that goal in mind, you define several personas. It will help you select the suitable data later. You write down the personas, their knowledge, and their interest on this project.
Can you classify your notes into the following audience personas?
Correctly classify the following examples as Human Resources Director, technical supervisor, or marketing staff.
Amazing job! You can now identify personas and understand what insights are useful to each of them. This can help you selecting the appropriate findings to convey the appropriate message that aligns with each stakeholder’s interests.
2.2. Showing relevant statistics
You have selected suitable data for your story on San Francisco salaries. Now, you evaluate which metrics you should use.
You want to convey the following message to the human resource team: “The total pay of employees has increased or decreased according to their job title from 2017 to 2018.”
You prepared the two visualizations below, but you are unsure which one is best.
One of the following options is True. Can you select which one is it?
- Both graphs showed clearly how the salary changed for the three different jobs. Both are meaningful ways of expressing the same insight and convey clearly the message.
- The graph on the left, showing total values, is the most suitable one. The changes in salary are observed. No adjustment is needed as you’d rather show raw than transform data.
- The graph on the right is the best way to convey the message. With percentage change, the magnitude of the salary change depending on the job is more evident.
That’s right! Adjusting the underlying data can help convey the key insight to the audience more effectively.
On a payroll
Good job on selecting the most impactful visualization! Your insight made an impact. The human resource team lead asked you to show more findings. You go back to your exploratory data analysis and select some data.
But you want to explore different variants of the same data to discover the best one for explaining your distinct insights to the human resource team.
Can you decide what data variant would be best suited depending on the finding you want to show?
Correctly classify the following examples as total values, change or ratio.
Well done! You now understand how you can select different data variations to make your story more attractive to your target audience.
It’s not significant
You have a big deadline ahead. You need to submit a report on the data analysis for the project on San Francisco salaries to your technical lead. He will read it and report your results to the senior management team.
You have a story, and you select data to support it. You want to show comparisons of average pay for people with different job titles.
You are hesitant to show p-values. You know that there are a lot of misconceptions around it. You decide to use it anyway. However, you plan to clarify any confusion about p-values, so your audience understands its meaning and trusts your results.
Can you classify these statements as good use or misuse of the p-value?
Correctly classify the following examples as either a good use or a misuse.
Well done! Always remember that p-value is not the sole indicator of a truth. You should evaluate carefully its values and consider alternatives.
2.3. Visualizations for different audiences
You are presenting your data analysis on San Francisco salaries to the business development department.
You have your story, and you select data and metrics to support it. But choosing the visualizations is still an ongoing task. You decide to speed it up by getting hands-on. You want to follow the best practices.
The message you want to convey is: “San Francisco salaries have been constantly going up in the last 4 years. The percentage variation is 10% annually. The number of people working in the private sector, such as software or biotechnology, have increased by 100k.”
Can you classify if these practices would be good or bad when presenting to the business department?
Correctly classify the following statements as being either true or false for choosing an effective visualization.
Amazing! Being able to select impactful visualizations for each audience is a key skill. Simple and engaging graphs go a long way in making your data story understandable.
Salary on demand
You have selected visualizations for the business development department.
It’s time to include them in your report and present them. You are aware of the steps you should follow to include and present visualizations effectively, and you want to do your best and impress your business coworkers. So you ask a colleague to help you organize your thoughts.
Can you order the steps for including and presenting visual data effectively?
Order the steps chronologically: the first step should be on top and the last step at the bottom.
Great job! By following these steps, you increase the chances of presenting your visualizations in an engaging and powerful manner.
2.4. Choosing the appropriate format
A communication problem
Your coworker has been working on a project on price predictions. He asks you to help him choose the most suitable format to deliver his results to the executive board as well as to his team.
You give him a set of advice and rules of thumb, so he can make an informed decision. When you arrive home, you realize that you made one mistake.
Which of the following advice should you not have provided?
- The amount of time the CEO can dedicate to getting up to speed with your analysis is an important factor in your choice of delivery format.
- If a software engineer in your team wants to continue your project with new data, the central piece of information to include in your meeting is the project conclusions.
- If your project manager, located in a different time zone, needs your results to communicate them to customers, a written report would be ideal.
That’s right! Your technical colleagues would be interested in understanding the methodology you used and the results you obtained. It will help them replicate your project.
Should we meet?
It’s Friday. Your project manager comes by your desk. She asks you about the status of your project on salaries for San Francisco employees. She tells you that you need to close the project. Fortunately, you have finished building the model.
But to close it, you need to communicate the results to the different stakeholders. After she talks with the people involved, you start to receive emails asking about the results. You need to decide if you are going to use an oral presentation or a written report.
Can you decide what type of format would be best suited depending on the situation and requirements?
Correctly classify the following inquires as suitable for an oral or written format.
Well done! Identifying the appropriate format is a key point in any communication strategy. Choosing the most suitable way to deliver your results can help your analysis have a bigger impact!
When in doubt
You manage to deliver the results to almost all the stakeholders. You are about to start writing the report for the founder when you get an email. Your founder is coming by the office the following Friday. Your manager wants to know if presenting the project during a meeting would be better.
You have second thoughts about changing the format. So you decide to write down beneficial and unfavorable aspects of an oral presentation.
Can you classify your thoughts correctly?
Correctly classify examples where an oral presentation is either beneficial or a unfavorable.
Well done! You should meet with the founder and follow up with a report you previously wrote for another stakeholder. Keep going and learn more about how to write reports in Chapter 3.
3. Structuring written reports
Now that you understand how to prepare for communicating findings, it’s time to learn how to structure your reports. You’ll also learn the importance of reproducibility (work smarter, not harder) and how to get to the point when describing your findings. You’ll then get to apply all you’ve learned to a real-world use case as you create a compelling report on credit risk.
3.1. Types of reports
Something to report
You need to present a report regarding your findings about customer churn and the predictive model you used, which you worked on in Chapter 1. Your project manager asks you to write it according to industry standards. You’re aware that this requires you to follow a strict structure. Your manager also specifies that the report will be shared with technical stakeholders.
First, you write the sections separately: it’s easier to handle that way. Then comes the time to bring all the sections together.
Can you organize the sections you wrote for the report in the correct order?
Order the report sections so that the first section should be on top and the last at the bottom.
Well done! You can now structure a written report. Always keep in mind that a well structured report is more likely to have an impact on your target audience.
Well done ordering the sections in your technical stakeholder report. Your project lead asks you to write a report to send to the directory board. They are non-technical stakeholders. You will revisit your previous report and tailor it as a summary report. But first, you want to refresh how a final report and a summary report differ.
Can you correctly classify the statements as characteristics of a final or summary report?
Correctly classify the examples as a feature of either a final or summary report.
Amazing! You should always tailor the report to your audience. A summary is usually written for decision-makers, so only key findings and recommendations should be included.
3.2. Reproducibility and references
Your manager asks you to write a report on your customer churn project for your peers at the New York office. She mentions that the team wants to replicate your work. After wrapping up the report, you add a link to your code repository. She looks confused and asks you why you did that.
You explain: If the New York team wants to replicate my work, then they should have access to the same ___ and the same ___ I used. However, if they want to achieve ___ , they can use their own set of tools.
Fill in the blank spaces by choosing the correct word combination from the options.
- data, code, replicability
- team, code, reproducibility
- data, code, reproducibility
- team, data, replicability
Well done! Reproducibility refers to the ability to obtain similar results that another team but using different data and code. Nowadays, there is a reproducibility crisis, so it’s very important that as a data science you follow best practices.
Your manager is very interested in learning more about reproducibility. She asks you to give a short presentation at the weekly meeting. You’re going to introduce the best practices to create reproducible data science results.
You prepare a slide presenting bad practices, and another one highlighting best practices.
Which of the following statements are considered best practices in reproducibility, and which should be avoided?
Correctly classify the statements as best or bad practices.
Good job! Ensuring reproducibility might make people confident with your work and use it to drive data-decision making. The way you write a report can contribute to this process.
3.3. Write precise and clear reports
Your coworker, whom you helped with a project on price predictions in Chapter 2, asks you now to read his report. He wants your feedback as the report is designed for the executive team. Particularly, you should focus on fixing any empty or fat phrase, or colloquialism.
Can you help him identify improvable sentences, and suggest a better alternative?
Correctly classify the examples as improvable or intelligible versions.
Well done! When writing and communicating results, it is fundamental to avoid sentences that can be confusing for the reader, and distract their attention.
Your coworker is very satisfied with your feedback. He now asks you to look at another report he wrote for technical stakeholders. You noticed that the writing style is not confusing and verbose. You decide to help him use more concrete nouns, and to fix redundant adjectives and run-on sentences.
Can you classify which of the following examples are intelligible, and which ones can be improved?
Correctly classify the examples as either the correct or uncorrected version.
Well done! Good technical writing should be direct, precise, and concise. To that aim, the right words should be selected to focus the reader’s attention on the facts presented.
3.4. Case study: report on credit risk
You are about to leave the office when you get a call from the operation director. She tells you that you need to write a report on the credit score to present to the advisory board. She explains: “They want to understand your analysis to help plan a strategy to pre-select customers for loans“.
Which of the following reports is the most suitable to write in this case?
- An analytic report including a heatmap with correlation between all predictive variables, technical details about your random forest model, and final recommendations.
- A short report showing a barplot with the financial risks involved if loans are given to defaulted and not defaulted customers.
- An analytic report with boxplots displaying the relationship between loan and customer traits, and a barplot of most important predictors.
Correct! An analytic report showing relationships between different variables is a powerful tool for data-driven strategies.
Report my credit
Your analytic report on the credit score project was a success. Your project manager was very satisfied with your work. Now, it’s time to write the summary report for the financial department director. You want her to understand the key findings. Particularly, that the most important features for predicting a loan default are income, age, and employment length.
Can you put in order the sections of your summary report?
Order the report sections so that the first section should be on top and the last at the bottom.
You have reached the end of Chapter 3. You now understand the best practices to write a precise and reproducible report for each audience persona. Keep going and learn how to build a compelling oral presentation in Chapter 4.
4. Building compelling oral presentations
You’ll finish by learning simple techniques to structure a presentation, communicate insights, and inspire your audience to take action. Lastly, you’ll learn how to improve your communication style and prepare to handle questions from your audience.
4.1. Planning an oral presentation
Is this the plan?
You need to present your findings on software development sector salaries for San Francisco employees and the predictive model you used, which you worked on in Chapter 2, to the executive team.
You open your computer and start drafting slides. But then, you recall the course you took on communicating with data. You feel that you are missing something. So you write down the facts you remember to build a compelling oral presentation.
Can you identify which of the following statements is correct?
Excellent job! Now you understand how to plan your presentation. It is a very important step towards building and delivering effective oral presentations.
An effective plan!
Well done! You identified which statements were correct. Now, you know how to plan your slides on salaries for San Francisco employees. It’s time to determine which sections you will include in your presentation.
Can you put in order the topics for the sections of your oral presentation?
Amazing! You have now planned your presentation outline! Jump in the next video to keep building the slides!
4.2. Building presentation slides
A color building
You have planned the presentation for your findings on software development sector salaries for San Francisco employees. Now, it’s time to build the slides. After putting your slide deck together, you ask your coworker to give you feedback. She makes comments on the different slides.
You start by reviewing the following graph slide.
One of the comments she wrote about the slide is not correct. Can you select which one it is?
- The headline looks stylish. Even though the spacing between the letters is small, the font adds originality to the slide.
- The amount of graphs displayed is adequate. The two full-size graphs can be properly seen and compared.
- It’s a good idea using the red square to emphasize one of the graph. It will make the audience focus on what graph you are explaining.
You are right! When choosing a font, you should consider their typesetting, as well as the spacing of letters and lines. They all affect readability. And you should choose a font that is easy to read.
Too much text
Your coworker also gives you feedback on slides that only have text. She feels like the slide can be improved a lot. She gives you some options on how you can improve the slide.
One of the following statement is not correct. Can you select which one is it?
- Instead of writing all the text in paragraphs, convert each concise main idea to bullet points.
- The headline is well formated using big letters. Also, it is specific and concise.
- Highlight the keywords in the text using red, green, yellow, blue, and magenta. Using several colors will help catch the audience’s attention.
- It would be a good idea to break the slide into two slides. In the first one, put the second paragraph in gray. Reveal the text in the second slide.
Correct! You should avoid using more than three colors in a slide as it can result in distraction. Also, you should be aware of color-blind people and avoid using red and green.
The right building
Well done building a text slide! Now, you need to review the rest of the comments from your coworker. She left you the feedback in two stacks of notes: one stack for best practices and one stack for what you should avoid. However, the stacks fell during your lunch break and are all mixed up.
Can you classify which of the following comments are best practice in building a presentation, and which ones should be avoided?
Correctly classify the comments as either best practices or practices to avoid.
Amazing! Designing slides efficiently is paramount when communicating your findings. A good story with a bad presentation can result in the message not being effectively received, and acted upon. Now, you’re aware of the best practices when building slides. Keep learning and find out how to deliver your presentation!
4.3. Delivering the presentation
Put it into practice
Excellent job creating those slides for the project on software development sector salaries for San Francisco employees!
Now that you have your slide deck ready, it’s time to deliver your presentation. Your project manager sends you an email. She asks you if you can have the meeting at the end of the afternoon. You are unsure because you need to practice. So you ask for some time for rehearsal. To convince your manager, you make a list of the advantages of practicing and rehearsing a presentation.
Can you decide which example you wrote is correct and which one is incorrect?
Correctly classify the statements as either true or false.
Great job! Practice, practice and more practice: one of the fundamental steps to give a good oral presentation!
Your rehearsal went very well! You are ready to present to the stakeholders.
However, you feel nervous. You know this is completely normal but you don’t want to look like you are unsure of your content. You take three big breaths and remember your training. There are some key points that will help you deliver the presentation effectively.
Can you identify which of this statements is a suitable point to consider, and which one is not?
Excellent job! Practicing is an excellent way to feel confident with your content. However, it’s normal to feel nervous. Following these guidelines can help you deliver your presentation successfully!
4.4. Avoiding common errors
The true mistake
Your coworker asks you to help him practice for his presentation. He wants your feedback so he can be prepared to deliver an important project to the company’s CEO.
He gives you a list of best practices he should observe during his speech. But you notice that some of them are not correct.
Can you identify which statements are wrong?
Correctly classify the statements as either true or false.
Well done! Understanding how to avoid common mistakes can help you deliver an effective oral presentation that eventually, can impact the decision-making process.
Do’s and don’ts
It’s the day before the presentation for your project on customer churn. Your team lead arrives and tells you that the CEO will be there to listen to your results. You start getting nervous.
You have planned your presentation. You built the slides focusing on a non-tech audience. You have practiced and rehearsed with some colleagues. But you really want to avoid making any common mistakes during your talk. So you decide to practice one more time.
To achieve your goal, you need to remember best practices so you can implement them. Also, you should make a list of common errors so you can avoid them.
Correctly classify the examples as either a best practice or a common mistake.
Congratulations! Now, you know the fundamentals of communicating with data. This is a powerful skill every data scientist should have, as it will help your projects have an impact on decision-making.
Amazing job! You’ve now mastered how to translate technical results to convey a clear and convincing story with data! Keep going and learn how to impact the decision-making process.Amazing job! You’ve now mastered how to translate technical results to convey a clear and convincing story with data! Keep going and learn how to impact the decision-making process.Amazing job! You’ve now mastered how to translate technical results to convey a clear and convincing story with data! Keep going and learn how to impact the decision-making process.