Highlighting your data science project

Foundationally, it is important for your reader to understand the question you were trying to answer, or the problem you were trying to solve. This frames the entire project. Without this information, the actions you took to answer the question will lack the context needed to demonstrate impact.  

When writing about the question at hand, consider the minimum amount of information your reader needs to know to effectively understand the situation. You do not want them bogged down in excessive amounts of detail; instead, help them get caught up to speed without boring them with superfluous information. 

Explain what your data source was. This can be illuminating for the reader, as it shows what you are comfortable working with, and what sets of data you already have experience extracting information from.  

Make sure to address two elements:  

1. What data was this source providing to you, and in what form? In addition, help your reader understand the quantity of data that you were sifting through. 

2. Why did you choose to work with this source? Were there any advantages or disadvantages to this choice?  

Describe the practical problems you encountered in your raw data and how you solved them. What issues did you face (missing values, duplicates, inconsistent formatting)? How much time did this take? Most importantly, explain the decisions you made about handling these problems. This shows you understand that real-world data is messy and that you can think critically about solutions. 

Explain the techniques or approaches you used to answer your question. You want to find a balance between providing technical details (what techniques or software did you use), and describing your work in understandable, plain terms.  

State what technical skills you utilized, and define the broader themes of your work (looking for patterns, making predictions, or comparing groups, for example). 

Share what your analysis revealed, using specific numbers and concrete insights rather than vague statements. Equally important is explaining how you validated your work. How did you check that your findings were reliable? Did you test against benchmarks, use different subsets of data, or measure accuracy? This shows you think critically about your results. 

Acknowledge the constraints in your project. These might be limitations in your data, your approach, or what you could conclude. Rather than viewing limitations as weaknesses, frame them realistically. That framing can help ground your work, and make you more appealing to employers. 

If you need help with framing your project, think of it as a case study with a beginning, middle, and end. In other words, what was the question being asked, what was done to answer the question, and what was the outcome or result of the work you did.  

Doing this has two primary benefits. First, it helps you provide structure to the project and gives you direction in writing about it. Second, and more importantly, it helps your reader see that structure and more easily understand the project. When they understand it, they are able to see the value and how you could apply those skills to their company or organization. 

Be considerate of the attention span your reader is likely to give your portfolio. How many projects are they likely to get through before they feel like they’ve seen enough?  

To that end, pick your best 3-5 projects to highlight, and focus on describing them in a robust manner (considering the checklist earlier in the article). Leave your reader feeling impressed with the projects that they see before leaving the page.  

Within the 3-5 projects you do select, try and display as much diversity in those experiences as you can. Diversity can mean a lot of different things in this context, so try to think outside the box and figure out each of your projects is showing off about your skillset and experience.  

Some different factors you could consider are the type of data used or the technical skills needed to work with the data. Also consider the project’s scope. Did you analyze the data or draw conclusions from your findings? Was the data presented out to anyone, and did it lead to any tangible outcomes? Help your reader understand how much of the data analysis process you had ownership over.  

Even if you had the most impressive data science project in the world, it might not matter if the way it is presented is not intuitive to understand. Consider whether the variables you are using will be easily understandable, or if the amount of information you are including is too dense. Bullet points and short paragraphs are preferable to long blocks of text. If you do have long blocks of text, consider including a summary for those looking to skim through the portfolio. 

Visual elements can also make your portfolio more engaging. Consider including some photos of data visualizations that you made, or something to represent the data set you are working with. Even if the visual is not of something strictly technical, it can help the reader stay engaged and interested in your work.