Building Your Data Science Portfolio With DataCamp Workspace (Part 2): Add an Analytics Workspace

Planning Your Analysis

It often pays off to plan your analysis before anything else. A good tip is to already create the text headers inside your new Python or R workspace to guide your analysis.

Here is a general structure that you can use, but feel free to adapt it to your needs!

Executive summary

An executive summary is a fantastic way to demonstrate your reporting skills. People don’t always have time to read every detail of your report. An executive summary provides the key information you want readers to take away from your report if they don’t have time to read it all.

This section might include the motivation or context of your project, the business decision you are trying to drive, the highlights of your findings, and your recommendations. Make sure that it is concise, clear, and impactful.

Introduction

Here is where you will introduce your analytics project. Questions you will want to answer are:

• What is the purpose of the analysis?
• What data will you be using?
• What techniques will you use to answer your question(s)?

Writing this section before you begin to analyze the data can be a great way to focus your efforts. It can also prevent you from wasting time answering questions unrelated to the goal of the project.

Methodology

This is where you will store all your work on importing and cleaning data, manipulation, relevant exploratory analyses, and statistical methods. Your methodology will depend highly on your data, research question, and skills you want to demonstrate.

Results

Here you present the main results of your analysis. This includes the results of any statistical tests you use and key visualizations that summarize your findings. Be sure to curate this to your project's motivation/problem statement.

Conclusion

In the conclusion, you should summarize your results for a non-technical audience and translate what it means in relation to your introduction. How do your findings relate to the questions/problem statements you posed at the beginning of your report?

The conclusion should also contain recommendations for future steps. For example, suppose your analysis convincingly shows that revenue has dropped since introducing a new pricing structure. What advice do you have for the product manager? It is more than just your job as an analyst to crunch numbers and share charts. You need to leverage your familiarity with the data to drive decisions and advocate for changes.

Appendix

Adding an appendix to a report can be very useful for keeping extra information, such as methodology and findings that are not directly related to the main analysis.

By including an appendix, you can ensure that interested readers can dive deeper while keeping your report focused to the most relevant findings.

Performing Your Analysis

Your methodology will differ depending on your chosen dataset and the goals of the project. Here are some tips to ensure your methodology section is as strong as possible.

• Brush up on DataCamp courses if you need a refresher on a particular topic. Our library of courses is incredibly comprehensive, and it is often easy to adapt code to your own purposes.
• Check out DataCamp Competitions for inspiration. We have held a wide variety of competitions = on the platform. By skimming through analyses of past winners, you may find inspiration for techniques and approaches for your own portfolio project.
• Make liberal use of the new AI features inside of Workspace. All Workspace cells now have a “Generate” option that allows you to add a prompt. A good prompt can quickly generate code and save you time. For example, using “Provide me with Python code to perform a left join on two pandas DataFrames with different column names.” will provide you with the syntax to merge two DataFrames without you needing to look up the specific parameter names. 

Instant joining syntax!

Polishing Your Project

Before finishing your project, you'll want to get it ready for a larger audience. Here are some tips to ensure it has the greatest impact.

Create great visualizations

Visualizations are a key part of any analysis. Here are a few key tips to remember when refining your charts:

• Keep it clear. Your visualizations should not require any additional explanation. It’s easy to overload charts with additional data irrelevant to the finding you are trying to convey. Your conclusion should jump out at readers, not be hidden amongst other data points and labels.
• Guide your readers. Make sure axis labels are clear, and use annotations where appropriate to draw readers’ attention. Another nice touch is to use the title to communicate the main finding and save the description of the chart as a subtitle. For example, “Shoe sales have risen since the introduction of our new vegan leather line” as your primary title, with “Shoe sales over time” as your subtitle.
• Use color effectively. Color can be a great way to highlight differences between groups and highlight key data points. But it can also become distracting. Be deliberate in your color usage. If you are using multiple visualizations with similar data points, maintain consistency. For example, if you use red to mark Canadian customers in a visualization, all future visualizations should use that same mapping.

A well-labeled plot from our Strava Data template!

DataCamp also has a fantastic course on the basics of data visualization, which covers best practices and when to use different chart types.

Tip: As mentioned previously, the Generate function inside Workspace can save you a ton of time looking up documentation for visualizations. For example, a prompt such as “Write me Plotly code to create a horizontal bar chart with the plotly_white theme.” inside of a Python code cell will generate a basic bar chart with a special theme that you can then adapt to your data!

Edit, edit, and edit

Make sure that your written work is concise, clear, and does not take more wording than is necessary. A long-winded paragraph that could have been replaced with a small table or chart can quickly lose readers.

The “Generate” prompt is also available on text cells, so you can, for example, ask it to “Proofread this paragraph and make it clearer and easier to read.” This can yield some great suggestions to improve your written work. Alternatively, you can ask a friend or colleague to read through it and give feedback on clarity, structure, and content.

Editing also applies to your code. Those interested in following on a technical level should have no problem reading through your process. Again, using the Workspace “Generate” feature can be a major game-changer. A prompt such as “Format my code and add code comments” will help correct any code formatting errors and provide helpful code comments to assist readers with your report.

Messy code without appropriate code comments

Cleaned up code with appropriate code comments

Prepare your workspace for publication

We already covered the hide code and hide output features in Part 1 of this series. Here are the types of things you may want to hide prior to publishing your workspace.

• DataFrame outputs used for exploratory analysis. Your readers don’t need to see every variant of a DataFrame you worked on in your report.
• Package install output. In some cases, your audience may be curious about what packages you used, but it’s not necessary to show the entire install output.
• Most code, such as cleaning, manipulating, and visualizing data. Depending on your audience, you may want to leave some code in, but be critical in what you include.

Don’t worry about making your report less technical: so long as it has been shared with the readers, they can also navigate to the workspace version and view the cells you have hidden.

Next Steps

In the next article, we will cover another type of portfolio project: machine learning reports. However, we already encourage you to head over to Workspace and start a portfolio project using the suggestions in this article!

You can also get started now using this Python or R workspace template with the suggested structure already set up for you!

If you’re interested in diving deeper into creating portfolio projects, check out this engaging webinar from author Nick Singh.