What are the best data visualization projects to add to your portfolio?

Pick a BI tool, learn SQL and try to get the data to your BI tool and start working on a dashboard, if you are a fresher, you can pick any use case and if you are an experienced professional who wants to get into the field of Data Analytics and BI, try to work on some portfolio projects that resonates your previous experience as well..

Data visualization first starts with understanding the Grammar of Graphics. Some of the best books recommended for this are The Grammar of Graphics by Leland Wilkinson and Storytelling with Data: A Data Visualization Guide for Business Professionals by Cole Nussbaumer Knaflic. Also to make powerful visualization it's not always necessary to use Tableau or PowerBI, even your Excel and PowerPoint can outperform. "Tools are important but technique is everything".

Data visualisation is a process where we bring out the hidden insights from data using graphs. The whole idea is to know , specific visual ways to represent the data in terms of charts to extract information in shareable and coherent ways to reach targeted audience in most appropriate ways. Right language skills like SQL, Python/R or tools like Tableau, Power BI along with knowing what chart is most appropriate to identify actionable insights become the key.

Crafting a data viz project is easy when you are passionate about the story it tells. What are you passionate about? Once you have a few possible themes dive into Kaggle and find something that aligns.

Creating a standout data visualization portfolio starts with strategic project choices. Select topics aligned with your goals – industry-specific data, compelling narratives, or creative experimentation. Tailor each project to showcase your expertise effectively.

If the data size is smaller we can use spreadsheets to perform the majority of data cleaning and sanitizing processes. If the data size is larger then you can use bash scripting or python pandas (ipython) to open and do the basic cleaning and fixing the missing as well as extreme features.

Kaggle is great for getting data sets that are already in a relatively clean format. However, clean is not what you will face in the real world. Kaggle is great again as usability scores out of 10 and tags that can give you an indication of how clean that data is. Once you have that data you have various options. Push it through Power Query to clean your data within Excel UI. If you are feeling brave, roll up those sleeves and get to grips with Pandas Dataframes in Python. Are you more into databases? Slot your data into a SQL table and then use SQL to clean it up.

In the data visualization journey, securing and tidying your data is critical. Whether you pull from online resources, APIs, or personal surveys, the raw data always requires learning and grooming. The goal is to detect and correct errors, fill in missing pieces, and reshape the information into a coherent structure. You're forming the raw data into something clear and meaningful using tools such as Python, R, SQL, or Excel. Think of this phase as preparing your analysis for success.

In my view, data's beauty lies in its potential to reveal insights. To unlock this power, finding and cleaning the data is crucial. Sources vary – repositories, APIs, surveys – but not all data is immediately usable. Cleaning is akin to polishing a gem. Duplicates, missing values, outliers – these require removal. Errors necessitate correction. Transformation and aggregation refine the data's essence. Tools like Python, R, or SQL are the artisans that craft this process.

I would say cleaning data is the most important though sometimes tedious part before any other afterwards analysis work! Also, you need to make sure the data quality is good so that we would not waste our time and energy analyzing the data! “Garbage in, garbage out”

Once you have created your graphs in Python it used to be difficult to share what you built locally. Not any more! We have Google Colab, GitHub and my person favourite is Streamlit. Streamlit is a community and a Python library that allows you to take your visualisations and deploy them to the cloud in minutes using GitHub. You can then share that URL with the world.

In my view, selecting the right tools and techniques for data visualization is akin to a craftsman choosing the right tools. Options range from simple charts to interactive dashboards. Factors like data type, project purpose, and audience guide this decision. Tools like Tableau, Power BI, D3.js, Plotly, Matplotlib are popular. It's a tailored approach to create impactful visual stories.

Build interactive dashboards using tools like Tableau, Power BI, or D3.js. You could create a dashboard that visualizes stock market data, social media metrics, or sales performance. This gives the audience a varied and interactive rapport about the project.

Types of variables in your dataset and what do you want to analyze (such as distribution, correlation, etc.) are two things to consider for you visualization. For example, if you want to show the relationship between elements over time, bar charts and line charts are your option. If you want to demonstrate the components or proportions between the elements of one whole, pie charts could be a tool. If you want to explore the relationships between large volumes of data, scatter plots, bubble plots, and box plots are all great tools.

Designing and refining visualizations, to me, is like molding clay into art. Applying data visualization principles, I choose charts, colors, and labels thoughtfully. Annotations, legends, and titles weave context. Aesthetics matter – contrast, alignment, and clarity speak volumes. Tools like Canva empower this creative journey.

As for visualization design and refinement, I think knowing your audience is the key. An effective visualization of data should be simple and relevant to its intended audience and be able to help your audience digest the information in a short period of time! For example, if you would like to build a dashboard to help senior leaders make decisions, Power BI is a great tool for this purpose. Through those self-service interactive data pages, senior leadership could drill down further into the data on their own. In addition, those reports from the dashboard can easily be transformed into presentation materials if people would like to use it for other sharing purposes

Sharing the projects you are working on is a great way to learn! You want things to look their best before sharing with others, so you put more time into making it the best it can be. Not to mention all that you can learn from the feedback you get.

Geospatial Visualizations, Time-Series Analysis and Network Analysis can be very helpful in Data Visualization. It helps to visualize with more insight in economic regional disparities, stock price movement and network data.

Sharing your work allows for refinement and enhancement, ensuring it's the best it can be. The more feedback you get can enhance the beauty of the data visualization you made. The insights gained from constructive feedback are invaluable, driving continuous improvement.

If you have experience with machine learning, build visualizations to explain model predictions. This can be particularly useful for decision-makers who need to understand model outputs.

I like the idea and want to share the words from Nancy Duarte, “Don’t project the idea that you’re showing a chart. Project the idea that you’re showing a reflection of human activity, of things people did to make a line go up and down. It’s not ‘Here are our Q3 financial results,’ it’s ‘Here’s where we missed our targets.’” This idea is so true to keep in mind!