Interviewing AI Assistants for Data Visualization

In today’s world, you need to run very fast just to stay in place. Technology is developing at an incredible speed, but I don’t believe it will replace specialists! I believe it will become a loyal assistant, helping to eliminate routine tasks.

Our team is keeping a close eye on all the latest AI innovations that could be useful for data visualization specialists, BI analysts working with data, graphs, and dashboards. These professionals are expected to deliver insights and visual representations of all kinds of data for various business needs. So much to do, so much to handle! A little help from an assistant certainly wouldn’t hurt.

We regularly review interesting AI tools, and in this article, we want to briefly introduce a few of them, while diving deeper into one of the most exciting ones!

So, months of researching the AI data visualization market have brought us the following insights! Our main tester has been Anya, our marketing director and neural network specialist! You can read a few of her articles on this topic on Medium, where she reviews some of these products. Highly informative reading!

Now, let’s take a look at the list of potential assistants! Who will we hire for our team? Tell us a little about yourselves, dear candidates. I’ve heard that some of you are great at working with data, but not all of you understand charts. And some can even build dashboards? How about a trial period and a test task? All agreed?

Perplexity
What it does: helps gather and analyze information
Drawbacks: the service may be too compliant with your request. Better phrase it as: “I want to understand whether small businesses need content marketing. Give me answer with pros and cons”

Athenic
What it does: analyzes data and builds charts
Drawbacks: works only with one sheet of data and sometimes makes odd calculations

Julius
What it does: powerful data analytics with a user-friendly interface
Note: Ability to build and customize charts directly in the service
Drawbacks: good for quick data analysis, not suitable for complex calculations or merging datasets

ChatGPT
What it does: analyzes complex data and prepares visualizations
Note: Ability to build and customize charts directly in the service through new queries
Drawbacks: always useful to check data additionally to see trends and key findings. Also, clarify the logic of its calculations.

Basedash
What it does: builds dashboards 100 times faster than manual assembly
Drawbacks: decent interface, but not the most convenient. Does not connect data from the Excel and basic tables but support many SQL databases.

Rows
What it does: replace traditional tables, simplifies data analysis, automates dashboard and report creation, and eases collaboration on projects
Drawbacks: advantages include easy data integration and availability of templates. But honestly, sometimes it’s simpler and more convenient to use good old Excel.

Polymer Search
What it does: to simplify the dashboard creation process and data visualization by automating their creation and offering intuitive templates and AI features
Drawbacks: currently the most interesting tool on the list. Give it data, and it will build a dashboard on it!

This last candidate intrigued me immensely, so I invited them for a second round of interviews and personally had several conversations with them. We worked together on data visualization tasks and dashboard building! Based on the trial period results, help me decide—should I hire this assistant full-time?

Trial period: testing Polymer Search

When you think about creating dashboards, what’s the first thing that comes to mind? Probably Power BI, Tableau, and a nervous twitch, because working with tables and charts can take hours. But what if there’s a way to do it faster, easier, and without yelling at your screen? Enter the neuro-analytics service—Polymer Search.

Why do we even need Polymer Search?

Polymer Search claims that building dashboards is now so fast, you won’t even have time to make yourself a cup of coffee. Mmm, with cream?  

Here’s what it promises:  
– Dashboard created in one click  
– No coding required  
– Automatic data visualizations  

Okay, sounds cool, but we’re here to test claims against reality. So let’s see how it actually works.

But first, our seasoned expert, ChatGPT, will help the newcomer get up to speed and provide the initial data!

Step one: Using GPT to generate the data

Before we begin, let’s prepare the data for analysis. In our case, it’s webinar data where we want to understand which lecturer generates the most profit and who is working ineffectively. We sent a request to GPT, asking it to calculate the ROI for the webinars. GPT provided formulas and even suggested a table template that we could use for further filling.

We won’t dwell on working with ChatGPT in detail, as much has already been written about it, and we’ve previously explored the various useful aspects of this tool for data visualization people in the article: Creating a Dashboard Using ChatGPT.

First request:

And the template:

Now we can give the data to our candidate!

Step two: Importing data into Polymer Search

Now that we have the data in hand, let’s start testing our main candidate! We’ll upload the data into Polymer Search. Let’s explore how to work with it effectively. After all, every specialist requires a tailored approach!

And here’s what impressed me right away: Polymer Search generated the dashboard almost instantly. What usually takes several hours (or days if Excel decides to corrupt the file) is done in just a few seconds here.

Of course, the chart formatting needs improvement—diagonal text on bar charts. The pie chart also needs some adjustments! But it’s a good start.

The boss’s heart is already rejoicing; it seems this candidate could be useful! And Polymer Search not only created a dashboard layout but also suggested several key metrics for visualization:

• Total revenue from webinars—it’s always interesting to see the overall numbers.  
• Average ROI by lecturers—who sells and who just talks. This is important in business…  
• Conversion of participants to paying customers—a metric that reveals the true state of affairs in the online school.  
• Quality assessment of webinars—to gauge how much the audience enjoyed it, as we are in it for the long haul.

Step three: Setting up visualizations

Working with visualizations in Polymer Search is also a pleasure. You can choose from various types of charts, from simple bar graphs to more complex diagrams. All of this is done in just a few clicks; it feels like you have a reliable and understanding assistant who knows what you want with just a half-word or a glance. Like a dream!

But let’s return to the real analysis tasks:  

For example, you can look at the revenue by webinar topics, which helps understand which topics have “hit the mark.” Or analyze the average ROI by lecturers to find out who brings in the most sales.

Interesting plus: Predictive data  

One of the coolest features in Polymer Search is the ability to forecast data based on what has already been uploaded. The tool doesn’t just visualize current data; it also attempts to predict what will happen next. 

For example, you might see that one of your webinars could grow by 65% in the coming months. However, I can’t say for sure that this is a reliable forecast to depend on since I haven’t seen their calculations. 

But it’s intriguing! The assistant doesn’t just mindlessly fulfill your requests; it also knows how to dream about the future!

Step four: Automatic data refresh  

Another advantage of Polymer Search is its automatic data refresh feature. When you add a new webinar to the table, the data on the dashboard updates immediately. No more struggles with manual refreshes; everything happens quickly and smoothly. This truly makes life easier, especially when the data changes frequently.

Downsides of Polymer Search

Of course, despite all its advantages, Polymer Search has some limitations that should be considered. No tool is perfect; it’s essential to understand your assistant’s constraints from the start and not overload it with tasks that are beyond its current capabilities.

• Data must be flat at input: The tool works only with a single table. If your data is spread across multiple tables, you’ll need to combine them.
• Not all visualizations hit the mark: Some graphs look good but don’t always provide accurate or useful insights. You might occasionally need to make manual adjustments.
• Limited chart customization: You can’t configure every aspect of the visualizations as flexibly as in Power BI or Tableau. This is sufficient for basic needs, but if you require in-depth control, you’ll encounter limitations.
• Inflexible grid: Moving elements around on the dashboard isn’t always convenient. The tool offers minimal customization options for object placement.
• Paid system: Yes, Polymer Search is not a free tool, but if you need to quickly create something decent and accessible online, the cost is justified.

Results of the trial period!

Polymer Search is a powerful tool for those looking to speed up the dashboard creation process and save time. Its predictive data and automatic updates make working with dashboards more efficient, but it’s important to be aware of its limitations. If you need to quickly create something on the fly, without delving deeply into settings and programming, Polymer Search is a nice choice for a personal assistant.

So, if you want to eliminate the routine in dashboard creation, Polymer Search is a good option. The time you save can be spent on something more enjoyable. 

Maybe finally brew that coffee? Mmm, with cream!

Well, that’s the overview of AI tools I prepared for you! 

If you have interesting ideas on how to use them effectively, share them on social media—I’d love to learn something new! The more wonderful assistants there are among AI tools, the more enjoyable the work will be for specialists! And I don’t believe they will be left without jobs; instead, many routine tasks can be shifted to the reliable shoulders of AI colleagues.