AI Lab projects
Learn about breakthrough AI innovation with hands-on labs, code resources, and deep dives.
Datamations
Data visualizations are an important part of data storytelling but sometimes they tell just a part of the whole story. Datamations help bring data to life by making it easy to create animations that show the process behind a plot or table.
Try the Datamations demo
The need
Informed decision making often depends on understanding the steps taken to arrive at a plot. Visualizations alone don’t tell a complete story of how data were analyzed, but communicating each step of an analysis pipeline can be time consuming and difficult.
The idea
We wanted to help readers by showing them each step of the data analysis process in an easy-to-understand visual format. Datamations are powerful animations that show the story of how a plot or table came to be.
The solution
We developed a tool for automatically creating animations from your data analysis code to show the story behind the final plots. Our research shows that these Datamations improve people's comprehension, and that people also prefer seeing them to static figures alone.
Making data clearer with animations
Creating a datamation doesn’t require that you change your code. Simply call an extra function from the Datamations library and run your code. Learn how the Datamations library can make it as easy to animate data as it is to analyze data.
Datamations creation process
Our Library uses a parser to classify the data values and analysis verbs in your code. It then translates the classified code grammar into key frames of an animation; values become visual states and verbs become animated transitions. Finally, a renderer on the front-end strings the graphics together to produce a datamation.
Technical details for Datamations
The Datamations library automatically turn code for a data analysis pipeline into plot and table animations. This process starts by mapping data values to states and operations performed on the data to transitions. The states, for intermediate and final values, are presented as key visual frames that are woven together by animated transitions. If Datamations, for example, identifies a “group by” operation in the code, it will depict the dataset being split into different groups.
Our “data value to states” approach is designed with different considerations for plots and tables. For instance, plots leverage natural frequency encodings when possible, and tables are shown in a “tidy” format, where each row corresponds to an observation and each column to a variable.
Each Datamation is modularly composed based on the data operation that corresponds to each step in the data analysis pipeline.
Resources:
For more details about our tool and research experiment, visit our Datamations research page.
CoModGAN
CoModGAN uses AI to complete an image that is missing significant amounts of visual information. Two neural networks work together to generate and verify a completed image.
Sketch2Code
Sketch2Code converts hand-written drawings to HTML prototypes. Designers share ideas on a whiteboard, then changes are shown instantly in the browser—helping improve collaboration between the designer, developer, and customer.
Innovation Tech Hub
Explore insights and behind-the-scenes technology for breakthrough AI innovations. From Tech Minutes videos to Technology Deep Dives, learn about the engineering that powers the future of AI.
AI School
Learn to create your own AI experiences with courses in AI technology. Engage with learning paths in conversational AI, machine learning, AI for devices, cognitive services, autonomous systems, AI business strategies, and responsible AI.
AI platform
Start building AI solutions with powerful tools and services. Microsoft AI is a robust framework for developing AI solutions in conversational AI, machine learning, data sciences, robotics, IoT, and more.

