AnalyticsCreator Congress 2022 | Power BI Embedded | by BI SAMURAI

In this session, I will talk about embedded analytics and the opportunities it creates for integrating employees, partners, and suppliers into business processes.

Before I go into the demo, I will briefly explain what embedded analytics means, because not everyone may be familiar with the concept.

My name is Vladimir Hansa. I have worked as a Power BI developer for more than five years and as a data analyst in the financial industry. Most recently, I worked at a trading market maker in Amsterdam before deciding to build my own business.

I am also part of BI Samurai. We are partners with AnalyticsCreator and focus on Power BI services, including building and implementing Power BI models and dashboards. We also provide complementary services around the Microsoft Power Platform, including Power Apps and Power Automate.

Together with Pascal from BI Samurai, I co-founded EmbedC. In this session, I will show the Power BI Embedded solution we created through EmbedC.

Gartner defines embedded analytics as a digital workplace capability where data analysis happens within the user’s natural workflow, without requiring the user to switch to another application.

In practical terms, this means analytics no longer sits only in a separate BI tool such as Power BI, Tableau, or Looker. Instead, it is integrated directly into the application people already use every day.

For example, if you work in a CRM system, the analytics you need can appear inside that CRM system, directly where you need it.

There are different levels of analytics integration. The most common approach is to use two separate applications. A company may work in an ERP, CRM, HR, or other business application, but when users need analytics, they export data, connect through an API, move data into a warehouse, or open a separate BI tool.

The next level is a reporting section inside the application. Larger applications often provide their own dashboard or reporting area. This can be useful, but it is still often separated from the actual workflow.

The strongest approach is analytics directly inside the page or workflow. This is what I will demonstrate. The user works inside the application and immediately sees analytics related to the data or action in front of them. This is also called contextual analytics because the insights are connected to the specific page, record, or task the user is working on.

If you are looking at an order in a CRM system, you might immediately want to see analytics for that customer: whether they have ordered before, their average order amount, and other relevant information.

This gives the user context and helps them make better decisions while they are working, instead of forcing them to leave the application and search for answers elsewhere.

We developed this demo to show how embedded analytics can work in practice. It is designed for businesses that have built their own SaaS products or platforms.

These companies often onboard clients into their application, but they may not have the developers, time, or budget to build a strong analytics section themselves.

One example is a platform for franchise management. A client may manage tens or hundreds of units in one system, but the platform may not provide advanced analytics. In that case, we can build an analytics extension, embed it into the application, and give their clients access to Power BI-based analytics without requiring a large internal development effort.

We also work with platforms that already have some analytics, but where the reporting is too basic. Many tools offer simple pie charts or bar charts, but customers often need deeper insights.

This is why we use Power BI Embedded. Power BI gives us strong capabilities for designing visuals and creating reports that may look simple but can provide complex and useful insights.

For this demo, we use market data. On the EmbedC website, there is a stock market demo that tracks the last 10 years of data for more than 8,000 tickers traded on NASDAQ.

We chose market data because it is updated daily and is easy for many people to understand. However, the solution itself is not limited to financial data. It can be applied to many industries and business models.

In the demo, imagine that the screen is any kind of application: a CRM, ERP, or another business application. The user workflow in this example is to manage and follow a stock portfolio.

Everything in the application, except the embedded analytics section, is built outside Power BI. When I change a position in the application, the analytics updates immediately.

For example, if I change the number of Autodesk stocks from two to five, that change is reflected in the embedded Power BI report. The values update, and the analytics recalculates based on the new portfolio position.

I can also add a new ticker. For example, I can search for Ford, add it to the portfolio, enter the purchase date, price, and number of shares, and then the analytics updates again.

The new ticker is added to the portfolio view, and Power BI immediately calculates the gain, loss, and related metrics.

This demonstrates the key idea of embedded analytics: the user takes an action in the application and immediately sees the analytical impact of that action in the same workflow.

The demo also includes more advanced visualisations. For example, there is a line chart showing the performance of multiple tickers over time.

Normally, a chart like this can become messy very quickly, but we developed a custom Power BI visual that makes it easier to explore. Even if the user cannot clearly read every line, the table and hover interaction make it possible to identify which ticker performed best over a selected period.

We can also show broader trends in the market, compare ticker performance, and create other insights based on the available data.

Another part of the demo allows users to compare ticker performance over a selected period. For example, I can view all tickers in my portfolio over the last 36 months and compare their gains or losses.

I can also select Microsoft and compare its performance against other companies. The visual makes it easy to compare each ticker against the selected benchmark and against the overall median.

We can also look beyond the user’s portfolio and analyse all tickers in the database. Since the demo includes more than 8,000 NASDAQ tickers, we can ask questions such as which tickers increased by more than 50 percent over the last 36 months.

There is one important disclaimer. The demo uses free data from Yahoo Finance, so some extreme values may not be fully accurate. Well-known companies are usually reliable, but unusual spikes should be interpreted carefully.

The purpose of the demo is not to provide financial advice. It is to show what embedded analytics can do.

The architecture is hosted in Azure. The backend includes a database that acts as the data layer for the demo.

In a real project, this could be a small database or a larger data warehouse, depending on the requirements. AnalyticsCreator could also be used to build and manage that data warehouse.

For data production, we use Python scripts to pull data from Yahoo Finance, which is publicly available. The data is ingested into the backend every night. We also use Power BI Embedded capacity hosted in Azure.

One benefit of Power BI Embedded is the pricing model. Instead of requiring a Power BI licence for every user, you purchase capacity.

The cheapest capacity is roughly one dollar per hour, or around 700 dollars per month if it runs continuously. From a certain scale, this can make sense because the same capacity can support many users.

Another benefit is that companies do not need to spend internal development time building a full analytics solution from scratch. They can focus on improving their core SaaS product while adding analytics as a valuable extension.

Embedded analytics can also create new revenue opportunities. If a company provides strong analytics inside its application, it can offer this as a premium feature or use it to increase the value of its product.

A further benefit is customer retention. If users can analyse their data inside the application, they are less likely to export the data into another tool. This makes the application more useful and encourages users to spend more time inside it.

Embedded analytics also makes insights available to more people. In many companies, analytics is mainly used by executives such as the CEO, CTO, or CFO.

With embedded analytics, operational users can access insights directly in the context of their daily work. That helps decisions happen earlier, before issues or opportunities have to be escalated.

In the architecture, Power BI is used to build the semantic model, measures, and reports. The application itself can be built with different technologies, because Power BI Embedded is largely technology-agnostic.

Power BI visuals can be embedded into different parts of the application and can communicate with the app experience.

That is the core idea: the analytics becomes part of the application, not a separate destination. Users can work with the application and interact with the data in one connected experience.

Thank you for listening. I hope this was interesting, and I am happy to answer any questions.

You can also reach us later through LinkedIn, our website, or EmbedC.