Intro to 3 Data Viz Tools: Matplotlib, Seaborn, and Plotly

Table of Contents

• Meet the Expert: Marwan Kashef
• What is Data Visualization?
• 3 Data Visualization Tools: Matplotlib, Seaborn, and Plotly
• 1. Matplotlib
• 2. Seaborn
• 3. Plotly
• How to Learn Data Visualization Tools

Data visualization is key to understanding patterns in data, discovering actionable insights, and sharing data with non-technical team members. With so many data visualization tools out there now, Lighthouse Labs Instructor Marwan Kashef walks us through Matplotlib, Seaborn, and Plotly, three top data viz tools relied on by today’s data professionals. Find out the pros and cons of using Matplotlib, Seaborn, Plotly, and the importance of understanding the basics like simple graphs and charts. Plus, Marwan shares how Lighthouse Labs is helping to teach data visualization to anyone who wants to add this valuable skillset to their toolbox! 

Meet the Expert: Marwan Kashef

• Marwan has a Master’s degree in AI, and has worked in data analytics for over five years.
• Marwan graduated from Lighthouse Labs bootcamp before returning to become an instructor!

What is Data Visualization?

Data visualization is an accessible way to understand patterns, outliers, and trends in data. Data professionals use many different types of data visualization, such as charts, diagrams, and maps as well as countless other diagram techniques.

How does data visualization help people understand technical information and results?

Data visualization adds a lot of color to discussions that were previously based on pure instinct. Visualizations typically differentiate instinct from fact in a way that is unbiased and neutral as long as the methodology and data are accurate. 

Data visualization is best used when you want to explain key trends and findings in a way that enables observation and discussion amongst your audience. Instead of a data professional simply stating the data’s conclusion, data visualization empowers the audience and the user to come to that conclusion on their own. Data professionals can then add their own supplementary commentary analysis. You want conclusions from the data to be implied by the data visualization, similar to how you don’t want to have to explain a joke. 

On a technical team, who is doing data visualization?

Most data work is taken on by the data team, but creating the requirements for a data visualization project is a unified, cross-team effort. If you’re creating a data visualization or a set of dashboards for a team, you need to have their input. You need to know what they want to be able to see and analyze, then you take those requirements and break them down into parts to see what you can and can’t do. 

Data professionals may reach out to developers to enable the collection of missing data or the creation of new data.

Should today’s data professionals have a good handle on older versions of graphing or charting tools in addition to new data visualization tools?

Absolutely! Most of the time, the more simple the visualization is, the more accessible it is. When data visualization is done well, no one needs to explain what the visualization looks like or how it works. You don’t want a data visualization to be so complex that it takes away from the relevance of what you’re doing. This is why it’s good to establish a foundation and understanding of older means of data visualization, like graphs and charts. Given that we’re heading towards a world where data visualizations may be creating themselves, it’s more important to understand when and where to use certain types of visualizations rather than today’s newest data viz tools.

Simpler means of graphing are also a great go-to when we don’t know where to start with our data analysis. Sometimes doing a bar chart, line chart, or scatter plot can be the quickest way to gain intuitions on what to do with your data. Simple graphs and charts enable more complex visualizations and analytics. 

3 Data Visualization Tools: Matplotlib, Seaborn, and Plotly

1. Matplotlib

Matplotlib allows basic to advanced plotting. Matplotlib is extremely flexible and fluid, so you have no limitations on what you can create and how it looks. In some cases, you can even do a picture-in-picture type of visualization where you simultaneously display two visualizations. Matplotlib allows a user to add markers and change colors, and for some types of visualizations, you can enable interactions. Matplotlib allows users to create a trend, and this is a very similar format to older versions of Excel. 

Keep in mind, in order to create visualizations in Matplotlib, you will have to use quite a bit of Python coding. If you’re using Pandas, you can quickly plot data using Matplotlib. 

When is Matplotlib used?

Matplotlib is best used when:

1. You’re trying to do quick data analysis. You want to create a basic visualization, so you can quickly see a data trend. 
2. You want to create a more integrated and meticulous visualization. A great use case of Matplotlib is when you have many elements at the same time, but you need to customize it yourself. You can create your own code and include plenty of functions, so you have full handle on what is actually visualized at the end.

Pros and Cons of Matplotlib

Pros:

• Versatile and accessible
• Customizable
• Good documentation 
• Universal tool that plugs into many back ends

Cons:

• Steep learning curve
• Users need to know Python
• Users need to understand the syntax of Matplotlib, which is based on the software, MATLAB

2. Seaborn

Seaborn is built on top of Matplotlib, so it’s technically an add-on. If you’re not familiar with Matplotlib, you can use just Seaborn for simple tasks, but it’s typically best to use both Seaborn and Matplotlib.

Compared to Matplotlib, data visualizations created in Seaborn look crisper and more aesthetically pleasing. Creating visualizations in Seaborn is straightforward; you won’t need to use as much code to create complex views. With just five or six lines of code, you can easily change the color scheme, and you can create views that are slightly complex without the need for large amounts of functions. 

When is Seaborn used?

Seaborn is used when you want to quickly create simple visualizations. 

Pros and Cons of Seaborn

Pros:

• Quickly creates simple visualizations
• Aesthetically pleasing visualizations

Cons:

• More limitations on what you can customize
• Visualizations are not as interactive and more static
• May need to simultaneously use Matplotlib to address issues in Seaborn

3. Plotly

When it comes to using Python, Plotly is one of the most popular data visualization tools. By using Plotly, a static diagram becomes fully interactive; you can switch between viewing single or multiple data points at the same time. This interactiveness requires only one piece of code, which is very powerful. You can create the same types of functions as in Matplotlib and Seaborn, but in Plotly, you have the luxury of being able to jump into details much more easily and creating histograms. Depending on the use case, you can create complex, interactive diagrams which makes Plotly more powerful than just using Matplotlib or Seaborn. 

Since all of the source code is Python, Plotly is fully customizable. This enables the abilities of Matplotlib in terms of customization while giving you the aesthetically pleasing portion of Seaborn. The beauty of Plotly is that it’s more than just creating data visualizations within a notebook — Plotly allows you to create interactive dashboards that can be used by anyone in your organization, whether or not they know Python. Users can see multiple visualizations within the same view, which enables them to answer deeper questions without having to recreate views. 

When is Plotly used?

Plotly is used when you want to make interactive visualizations and dashboards. Anyone can log in on your private server and see updated information captured within the visualization. Plotly gives more accessibility to data and allows for more richness in conversation.

Pros and Cons of Plotly

Pros:

• Accessible and creates aesthetically pleasing visualizations
• Easy to create interactive components in visualizations

Cons:

• Steep learning curve
• Users need to know Python
• Uses its own syntax

How to Learn Data Visualization Tools

Lighthouse Labs offers online data analytics courses purposefully designed for data beginners. These upskill courses provide learners with an understanding of data analysis and industry tools, and the curriculum covers what is data science, how data is collected and created, and how to leverage and visualize data to derive actionable insights. These courses begin with Excel to get learners acquainted with the concepts of data visualization, and then we branch out to other data visualizations tools, like Tableau.

Who are the instructors for the data analytics upskill courses at Lighthouse Labs?

Instructors for the data science courses at Lighthouse Labs include anyone who has worked in the industry, worked in education and data science, or has pivoted into the industry. For those instructors who have made a career change into data, we have an appreciation for the difficulty in the data science journey, especially when it comes to changing industries.

What can data professionals learn from failed data visualizations?

If given data is deemed as being ineffective or failed, a professional can take that feedback into account and create a custom visualization to address those concerns. These can be issues like how the data visualization was communicated, if it was too cluttered, or if there was too much cognitive overload.The solution can be to make a visualization more simplistic or complex, depending on the issue. If there’s confusion about the best practice, you can use a process of elimination by showing multiple views from the same data set. This allows you to see which set the audience absorbs efficiently. 

As a data professional, I keep a checklist that incorporates feedback from my past projects. For example, avoiding lime green or deep red in visualizations so your visualizations are colorblind-friendly. It’s important that you improve accessibility and inclusion in your data visualizations.

What tools or methodologies should data professionals learn in order to stay relevant?

It’s important to master your foundations, like understanding Python, Excel, and SQL. If you can yield the power of leveraging databases and cleaning them properly, you’re a data rockstar already! Understanding Matplotlib, Seaborn, and Plotly will be a great asset, especially when you’re asked to create a report based on data analysis. It’s also good to be able to yield either of the industry tools, PowerBI or Tableau.

What are your top 3 resources for anyone launching a data career?

1. Towards Data Science is a comprehensive resource with a strong community that’s ever-growing. They post articles on new practices and algorithms, as well as guides on how to break into data science, apply to different data jobs, and prepare for interviews. If you’re working on a project or you encounter an error, Towards Data Science has created a walkthrough or codealong you can reference!
2. Infoworld.com posts trends and key patterns that are happening within the tech industry now. 
3. TechCrunch highlights how data analytics, data science, and AI are actually being used within the tech industry today. 

Find out more and read Lighthouse Labs reviews on Course Report. This article was produced by the Course Report team in partnership with Lighthouse Labs.