Analytics
News, events, thought leadership and more.
Some of the most exciting product news tends to revolve around new releases and new partnerships, and we are thrilled to finally be able to share more details about our recent news regarding our partnership with Plotly.
In Alteryx Analytics 11.0, we introduced Plotly's open-source visualization technology in several predictive tools within the Alteryx Designer. With our recent partnership announcement, we are deepening our relationship to deliver in-line visualytics - a visual representations of data as it moves throughout the analytics pipeline – to the data cleansing, prepping, blending and modeling process.
To further explain the partnership and vision of Ploly and Alteryx, we recently sat down with Matt Sundquist, co-founder of Plotly, Ashley Kramer Alteryx’s VP of Product Management, and Katie Haralson, Alteryx Product Manager of in-line visualytics to ask them a few questions.
In this first blog interview, of our three-part series, we sat down with Matt Sundquist, COO and co-founder of Plotly, and asked him to share a bit more about himself, his company and what makes Plotly so different.
Tell our audience a bit more about Plotly and what inspired you to create it?
The inspiration behind Plotly is simple: we know that data science tools are powerful, yet they’re often complicated, offline and not reproducible – basically, they’re painful. Most tools make it challenging to combine data, analysis, R, Python, databases, graphs, folders, varied file types, discussions, revisions, and spreadsheets to create something meaningful. It’s complicated and challenging for teams to work together even where analytic understanding is strong, since technical and coding experience and coding languages vary.
Let me give you an example. I became interested in the data analysis workflow while doing a research project about the Supreme Court, where my collaborators and I spent months analyzing and communicating our results. Our process looked something like this:
- Collect and access data with Python
- Store in a database
- Query data with SQL
- Transform and clean the data with R
- Explore the data with Excel
- Visualize the data with D3.js and R
- Present in PowerPoint, website, and paper
- Share with email
This process is complicated and creates three difficulties for publishing results. First, creating interactive graphs is challenging. Second, as the code, data, graphs, discussion, and revisions aren’t saved or published together, people can’t edit the results in a collaborative online environment. Third, data scientists don’t want to build a website to publish data applications or graphs.
Because of challenges such as these, teams lose time and energy trying to collaborate, or often just don’t.
With Plotly we want to lower the barrier to entry for creating complex, collaborative data visualization, and help bring disparate sources, coding languages and teams together. We want to make the flow of information easy, particularly between teams of varying coding languages and technical abilities. This is why we feel strongly about partnering with Alteryx as we know that Alteryx has focused on lowering the barrier to entry and uniting the workflow on the data analytics side of the house.
How many people use Plotly today?
Matt Sundquist: Currently over four million users turn to Plotly, and that number is growing. Many of our users leverage Plotly.js, our JavaScript library, and our APIs for R, Python, and MATLAB to make interactive graphs and share data.
What makes Plotly different?
Matt Sundquist: I would say four general aspects. First, all of our plots are interactive, web-based, and editable with a GUI, which helps analysts to be productive right out of the gate.
Second, Plotly is interoperable: regardless of a user’s coding language or coding experience, a user can still collaborate and add to the same plot from any language, and edit the plot with or without code.
Third, Plotly is open source. That means anyone can copy and use our code in their own applications. We believe in using and creating open source code: it means we make better software, anyone can re-use our code in their applications, and that we get to partner with teams we’re excited about, like Alteryx.
Fourth, the graph, data, and code to make the graph are always together. All of the graphs (or plots as we call them) created using Plotly are hosted at their own URL - think of it as a central clearinghouse for the data - and from that URL, you can access the graph and data and embed them into PowerPoints, blogs, or dashboards, as we’ve done below. This helps drive collaboration and ensure work is reproducible.
Example Plot: https://plot.ly/~jackp/17843/
What kind of charts and graphs can Plot.ly create?
Matt Sundquist: Plot.ly supports dozens of chart types that address a multitude of needs - line plots, scatter plots, area charts, bar charts, error bars, box plots, histograms, heatmaps, subplots, multiple-axes, polar charts, bubble charts, shown below, and more.
Though statistical, financial, and scientific graphs are powerful tools, they can be difficult to make, so we want to make creating these charts easy, beautiful and interactive. That’s why Plotly allows users to make their own volcano plots, matrix charts, small multiples, slope charts, dendrograms, Manhattan plots, ternary plots, network graphs, shown below, and a few dozen more.
What were your three priorities when creating Plot.ly?
Matt Sundquist: We believe that everyone should have access to powerful tools for statistics and data science, and we want Plotly to be as robust as possible. That’s why our platform is open-source; our code can be copied, edited, and re-used, so anyone can contribute. We want to maintain collaboration, support and ensure the perpetual motion, existence, and development of the product. Being open source means Plotly as a product will always exist, and we can continue to deliver a cutting-edge offering for exploratory visualization.
The other area of focus for us is to make sure that we aren’t adding to the already sprawling ecosystem of tools and complex data science analysis workflows. As I mentioned earlier teams are already wasting time and energy trying to collaborate, or just aren't collaborating. This is a problem. To make the best decisions, everyone needs to be able to collaborate, have the right data, and have the best tools at the right time, which is why we want to make sure Plotly is as accessible as possible regardless of technical ability.
We want to ensure that it is easy to publish graphs and data together and make sure it’s reproducible and shareable. Once an interactive graph is ready, sharing is as easy as sharing a link. The graph, data, and code to make the graph are always published and available together on the author’s profile, and can be edited with others online.
Most important question of all, how did you come up with the Plotly name, and when do you use the dot in “Plot.ly” and when do you not?
Matt Sundquist We chose Plotly because this is the technical term many users use to describe a graph or chart. The word plot is the term in MATLAB for making graphs. The most popular static graphing library in Python is called matplotlib, because it is a mathematical plotting library. And the most popular static graphing library in R is called ggplot2, a plotting library based on the grammar of graphics. The term “Plotly” helps provide continuity across applications our technical users know.
Now, when it comes to the dot in Plotly, we typically use the dot when discussing the website. The .ly is because that’s the internet country code top-level domain for Libya. It is substantially cheaper to buy a domain with the .ly ending than to buy a .com domain, and when you start a company, lean budgets are often what you have to work with, and now it’s just part of our culture.
Up Next:
Stay tuned for our next blog interview with Ashley Kramer, VP of Product Management at Alteryx , where she explains how Alteryx + Plotly together are bringing inline visualytics to the data modeling process.
