Understanding Data Science in Adobe Experience Platform

[#1]

Authors: Richa Sharma, Manasvi Mahant, and Douglas Paton

In this post, we take a look at what data science is. We explore common misconceptions about what is and isn’t data science. And we talk about how it can be used to help solve business problems using Data Science Workspace for Adobe Experience Platform.

Data is at the heart of everything we do these days. Every action, decision, and interaction that we make on the internet results in massive amounts of data.

When we are able to harness this data, it gives us the ability to make more informed decisions, create products and offers that are backed up by data, and make connections that we’ve never been able to make before.

Data science helps us make sense of all this data. It allows us to understand where the data comes from, what it represents, and how it can be processed and recognized faster to derive meaningful insights from the data. These insights would be helpful for companies to make better business decisions and generate higher revenue. This gives you a competitive advantage over your competitors who aren’t using data science in their businesses.

Unlike business intelligence, which relies on the use of structured data to provide a better understanding of business operations, data science makes use of all the data that you may have at your disposal, both structured and unstructured, to not only better understand your business. But data science also makes it possible to make predictions based on that data, something business intelligence tools can’t do.

Data science hierarchy of needs

Data science provides us with a framework for working with the data that you accumulate through the course of doing business. The framework can be broken down into five stages, effectively creating a hierarchy of needs.

1. Collection — The first step is collecting up the data that you have in your business. It’s any and all touchpoints that customers have with your business and includes everything from website analytics to geolocation services.
2. Move/Store — Moving data from one point to another (using something like Adobe Experience Platform Pipeline) in anticipation of the next stage.
3. Explore/ Transform — The data is cleaned and transformed into a format that can be understood and manipulated by data scientists. In the Adobe world, this is Experience Data Model, or XDM.
4. Aggregate/Label — At this stage, the data starts to be sorted out and labeled as a way of getting it ready for use.
5. Learn/Optimize — The final stage before you do deep learning. This is where you make sure everything is running smoothly with your data to ensure that nothing happens that could potentially affect your results.

Figure 1: The Data Science Hierarchy of Needs

Data Science Workspace in Adobe Experience Platform

To manage data science within the Adobe ecosystem, we have Data Science Workspace. Data Science Workspace uses machine learning and artificial intelligence to unleash the insights that are locked within your data. Integrated into Adobe Experience Platform, Data Science Workspace helps you make predictions using your content and data assets across Adobe solutions.

Using Data Science Workspace you can make predictions using your content and data, create intelligent services — APIs powered by machine learning.

Data Science Workspace was designed to gain insights into business outcomes while considering the reliability, speed, security, and compliance.

Use Case

To help you better understand how this works in the real world, let’s take a look at how data science can help during a pandemic.

We have seen the COVID-19 outbreak around the globe and, in order to save lives, doctors and healthcare workers are working day and night tirelessly.

Let’s say during this epidemic 1,000 patients are admitted into a hospital and we have a team of 20 doctors. Each doctor is assigned to work with 50 patients. Due to the increased number of patients, it becomes difficult for healthcare workers to cater to all the patients seamlessly.

Doctors and Healthcare staff can use this information in a variety of ways. For example, they could prioritize patients who need immediate medical attention or identify patients who have improved enough to be discharged.

This reduces the load on doctors and healthcare staff. And allows them to save time for themselves and take care of their health in order to save other lives.

Adobe Experience Platform can bring together the data related to the patient's medical history, existing diseases, real-time data, vitals, and other tests. With Data Science Workspace we can deploy machine learning algorithms to analyze the data and provide valuable insights that can be shared with doctors and medical staff.

Step 1

Real-time data related to patients medical history, existing diseases, vitals, ventilator status, and other tests are collected. The result would be something like the sample CSV below.

Figure 2: A sample of patient CSV data

Step 2

After the data has been received, it gets ingested into Adobe Experience Platform and models can be built on the data using Data Science Workspace. You can visualize the results/analysis in the form of graphs also.

Figure 3: Data Science Hierarchy of Needs

Figure 4: Individual Patient Glucose Level graph details. Vertical Axis represents the glucose level and the horizontal axis represents the date.

Step 3

Hospital staff can then use the insights they received to send emails and push notifications to concerned doctors and medical staff who can use this information for their diagnosis. And, can attend patients which would require immediate attention.

Conclusion

Data science provides you with opportunities to drive key business decisions using the data that you have at your disposal. Data science provides a powerful way to identify new opportunities for your business that might not have been obvious without the help of data to make these connections.

Different businesses can benefit from the insights provided by using data science and it would not be wrong to say the future belongs to data science.

For a more comprehensive look at Data Science Workspace, feel free to explore our walkthrough documentation.

Follow the Adobe Experience Platform Community Blog for more developer stories and resources, and check out Adobe Developers on Twitter for the latest news and developer products. Sign up here for future Adobe Experience Platform Meetups.

References

1. Adobe Experience Platform — https://www.adobe.com/experience-platform.html
2. Adobe Experience Model (XDM) — https://www.adobe.io/apis/experienceplatform/home/xdm.html
3. Data Science Workspace — https://www.adobe.com/ca/experience-platform/data-science-workspace.html
4. Data Science Workspace Walkthrough — https://www.adobe.io/apis/experienceplatform/home/data-science-workspace/dsw-overview.html#!api-spec...

Reference for Data Science Workspace terminology

Algorithm : Standard DS Techniques such as classification, regression, k-means, and clustering.

Feature: An individual measurable property or characteristics of a phenomenon being observed.

Feature Engineering: Process of converting raw data into a usable form for analysis using domain knowledge.

Hyperparameters: High-level properties of a model, different from standard model parameters that are usually fixed like depth of decision tree, number of hidden layers, and learning rate.

Instance: An occurrence of the recipe configured with the right data definition to help solve specific biz problems — one recipe can create many instances.

Jupyter Notebook: An open-source web application that allows you to create and share documents that contain live code, equations, visualizations, and narrative text.

Recipe: A proprietary algorithm or an ensemble of algorithms to help solve specific business problems.

Service: Service that is created from a “trained model” to be used in building experiments.

Trained Model: An instance of the recipe that is trained using historical data to learn from. The trained model finds patterns in the training data to predict the target.

Originally published: Apr 30, 2020