Implementing embedded analytics is the easiest method of integrating analytics, since reports can be rendered with a simple copy-and-paste of the embed code. Embedded analytics can be found on public websites we interact with on a daily basis as a result of the ease with which reports can be embedded into a website, blog, or news article.
The use of embedded analytics on public websites has become more common as our data-driven culture continues to grow. Most likely, you have interacted with similar visualizations without realizing that they constitute basic embedded analytics.
These types of analytics implementation can benefit content creators who, in the past, may not have been considered data-savvy; they provide readers with a visual representation that their brains can more quickly interpret than text alone.
It is possible to integrate advanced embedded analytics into a third-party product or web portal. Most commonly, this is done by software-as-a-service (SaaS) companies with core applications that lack native reporting capabilities. By partnering with an analytics vendor who sells this software, a SaaS company can integrate the vendor's technology into their native SaaS product rather than develop their own analytics solution from scratch.
It is important to note that advanced embedded analytics implementations, from a technical perspective, require software engineering, but offer more robust functionality than basic implementations such as advanced filtering, row-level permissioning, near-real-time data calculation, and report scheduling. Here are come examples of embedded analytics you may have seen.
1. Onbrane
Onbrance produces innovative technology-based platforms for the debt market. They chose to partner with Toucan to bring customer-facing embedded analytics to Their website that could be used by everyone.
Onbrane’s users are able to have a comprehensive analytics dashboard rendered in real-time, highlighting important data that customers are able to leverage within their website.
The seamless integration between the two products allows users to access analytics without having to navigate between multiple systems or even know that the dashboard was created in Toucan’s entirely separate application.
2. Politico
This article in Politico compares a series of maps of the United States, each showing the degree to which gun ownership varies by demographic using embedded data visualization. Whenever you hover over any state on a map, it shows you the percentage of voters who (don't) possess guns and who they voted for.
With this visualization, you can understand how gun owners are divided along party lines at-a-glance by comparing the classic colors associated with the Democratic and Republican parties. The pop-out information you get when you hover your mouse is also great, since it adds a dynamic element to the map that otherwise would have been displayed as a series of 50-item graphs.
3. Zilliow
Zillow, an online real-estate company, displays a market value line graph for each house listing using embedded data visualization.
When the user navigates across the years listed on the x-axis, the displayed home value updates. With this embedded graph, users can immediately see how a home's value has changed over time, and interact with it.
4. Manhattan Marathon
In embedded data visualizations, the reader is often given a sense of how the data compares to what they know. Comparing a human figure to a bear displays the bear's size more effectively (for example, a human figure standing next to a bear).
This interactive is an excellent example of how data can be made accessible. When you enter an address, the tool will display a visual representation of a race and how quickly the participants ran it. In this way, the average Joe can gain a better understanding of how amazing these athletes are. For example, they might examine the marathon:
Then consider how long it would take them to walk, run, ride or drive even a portion of that distance and you will realize how the runners are. The cruising speed for these athletes - and that is not even halfway through the race - should be no surprise to anyone who has ever walked or jogged from Midtown to the Bronx.
It is the goal of a good data visualization to enable people to connect with statistics, and this feature accomplishes this effectively by providing a familiar framework for understanding statistics.
5. Disappearing shorelines
This interactive embedded data visualization illustrates how parts of cities may disappear from water without engineered protection. By moving the slider, different percentages of land in major US cities would be covered by water as sea levels rise. By increasing the tide, you will be able to see what landmarks and areas will be submerged as the tide rises. Readers can easily flip back and forth between different maps by overlaying them on top of each other.
6. Uber drivers and wait times
In this interactive embedded data visualization, you can see how the number of drivers idling nearby is having an effect on the wait times for passengers on a city street. There is an option for viewers to choose from 50-250 drivers nearby in order to see how the number of idling drivers and wait times change as time passes.
The interactive graphic does a really good job of giving an overview of how many cars have to be present for the wait time to be low, and what this means in practice. The above argument provides a much stronger support for the thesis than just saying you need 125 cars idling for an eight-minute wait time. This interactive data visualization is a great way to convey a message and make an impact.
7. Game of Thrones opinions
As illustrated in this NYT article, Good, Evil, Ugly, Beautiful: Help Us Make a "Game of Thrones" Chart, data can be a fun way for readers to interact. As part of the NYT's contest, readers are asked to plot characters on an axis from beautiful to ugly, evil to good. You will then be able to see the aggregated data gathered from all readers. (Alternatively, you may skip straight to the aggregate without plotting anything.)
This article is engaging because of its reliance on reader sentiment, which has been captured in an interactive embedded data visualization. Moreover, it provides a heat map showing the average distribution of characters, which illustrates that some characters have a much larger consensus than others, while some are deemed more beautiful than others.
A highly engaging interactive embedded data visualization can be created by testing readers' intuition and then revealing what more data reveals about the question you pose. You can drive deeper consideration of your content by asking readers to categorize their own thoughts before revealing your data.
8. Renting vs buying a house
By now, it should be obvious that interactive embedded data visualizations are more enjoyable to work with. Despite the fact that mortgages aren't everyone's idea of a leisure activity, the interactive slider on these graphs can make planning your financial future more enjoyable.
This is a great example of how adding interactive embedded charts can spice up otherwise straightforward information. A reason why it works so well here is that people are interested in playing with the term and rates, and creating a live representation of their thoughts is a great example of how it helps people decide whether to rent or buy.
9. N.F.L. playoff simulator
Sports simulators and prediction tools use the most amount of embedded data visualizations to convey information effectively to fans. The N.F.L playoff simulator calculates 174 duodecillion ways that each team could end its season.
Upon clicking on a percentage, you will be taken to a second page of interactive graphics that allow you to explore all the different ways your team could reach - or win - the playoffs.
Thanks to our modern digital capabilities, probability and decision trees are once again brought together. Interactive features like this are something only possible today, and boy are they fun. The combination of this chart with a more traditional chart is designed to bring things back down to earth and provide readers with an understanding of which scenarios are likely to favor their team.
10. Rio Olympics
Embedded data visualizations such as this one illustrating all Olympic medals won since 1896 are excellent examples of bubble charts. A glance at the dataviz provides a wealth of information regarding who has won the most medals. A pseudo-map helps viewers to visualize trends in the data - such as the fact that Central America and the Pacific islands have fewer medals than, for example, North America or Western Europe.
Hovering over each circle reveals more information, including the overall medal total and medals broken down by gold, silver, and bronze. Additionally, the colors are bold and serve to help users distinguish geographical regions.
11. School district comparison
Another excellent example of context and how interaction can enhance a reader's understanding is the map showing how school districts compare. This embedded data visualization plots how far ahead of their grade level students' test is in comparison to their parents’ socioeconomic status. Although this is a telling picture on its own, it becomes even more compelling as you scroll over the graph.
If you hover your mouse over a dot on the graph, you will see which school district it represents, how much above average the students test, the median family income, and the breakdown of ethnicity by percentage.
By including this interactive feature, this chart is able to resist the tendency to be reductive. As one scrolls over the peaks and troughs of the graph, it is evident that poorer students of color tend to make up the majority of the schools with the lowest performance, whereas richer white students tend to make up the majority of the schools with the highest performance. It is not surprising to anyone who is aware of the widening wealth gap between white and black citizens of the United States since the 1980s. The inclusion of this demographic breakdown in this chart enables this data to tell a more complete story as a result of interactions.
12. Historic trends in the Arctic’s sea ice
The environmental warriors aren't kidding when they say you'll see a trend. Charts in the article illustrate how the area of arctic sea ice has changed from 1979 to 2017, and just glancing at the chart alone is sufficient to demonstrate that the amount of ice is diminishing each year. The graph displays specific statistics for each year when you move your cursor over it.
With the increasingly hot pink hue backing up the trend of more and more ice melting, they are using an intuitive color scheme. Additionally, we like the fact that it shows the variation in ice coverage without simply stating that the difference is in millions of square kilometers - since that is not an easily visual metric.
Big takeaways
There are so many ways you can represent data in an interesting and interactive way. Let’s go over some common themes we saw in this article:
- Add value through interaction. There was more to the embedded data visualizations than just being fun. Interactivity allowed them to create charts that were otherwise difficult (or impossible) to plot and interpret. Others helped the user reflect critically on their own choices and perceptions. It is a waste of time to create a chart or graphic just for the sake of throwing it out there. Your visualizations should help users better understand your subject and discussion.
- They don’t have to be complex to be good. Your readers can gain insight from something as simple as a pop-up with more information. Take the school district comparison, for example. If you can’t build a full interactive model of the entire dataset, don’t let that hold you back from creating something interactive. You can be smart about any data you’re trying to present — if you’re dealing with geographic data, for example, pull out a map and ask yourself how you can use it to illustrate your point. If you’re doing political visualizations, let that pick your color scheme.
- They build an intuitive understanding. A good embedded data visualization should assist people in understanding and engaging with data more effectively. Whenever you are creating a visualization, keep your average user in mind. Maybe you are catering to experts in your field who do not require exhaustive descriptions of your data. As we saw in renting vs buying a house, if you are creating for a more general audience your users may need an additional key to help them understand the chart.
- Embrace the concept of data storytelling. Through storytelling, humans are able to learn and remember information. When adding an interactive visual, you have the opportunity to create a narrative based on the data. This is called data storytelling. Take Uber drives and wait times — instead of just listing the wait times, the reader builds the narrative of how and why the changes happen through the embedded data visualization.
The data represented here ranged from demographics to sports statistics to global economic systems. Using interactive embedded data visuals, any type of data can be explored, and you may have more data that can be visualized than you think - it does not have to be two coordinating points on two axes. The more creative and interesting ways you can present your data, the better.
