cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
Showing results for 
Search instead for 
Did you mean: 

Community Tip - If community subscription notifications are filling up your inbox you can set up a daily digest and get all your notifications in a single email. X

Make your Data Talk! Using Visualizations as an Interface for Intelligence: Webcast Replay + Q&A

No ratings

image

 

Based on real use cases and industry-leading solutions, this webinar looks at how developers can deliver valuable analytical outputs through experiences that ensure easy consumption of trusted analyses.

 

By taking a deep dive into a range of the analytics capabilities within ThingWorx, we will demonstrate how you can visualize complex analytic outputs to help your users understand what really matters in your data - and ultimately make quick, insightful business decisions.

 

Q&A

 

We didn’t have time to get to all of the questions during the live webcast, but we’ve answered them here on our blog. Have any additional questions? Please leave us a comment.

 

THIS SESSION TALKED ABOUT CONNECTIVITY WITH OTHER ANALYTICS PROGRAMS, SUCH AS R. HOW WOULD THINGWORX INTERACT WITH THE R PLATFORM?
There are multiple reasons for using R and other analytics tools. Say you’re using R to build a predictive model, for instance—you can interact with data and load that information into our Analytics Server, even outside Analytics Builder. You can also interface in script within ThingWorx to run certain calculations driven by R. Moreover, even if you are building a model in R or any other tool, you can use Analytics Manager to operationalize data coming in from a Thing model that’s being scored against variables created elsewhere. Our ultimate goal is not to migrate you away from the tools you’re already using, but rather augment the development experience with ThingWorx integration.

 

HOW WOULD A DEVELOPER GO ABOUT CREATING & OBTAINING JSON AND CSV DATA FOR ANALYSIS?
In terms of creating a historical data view, there are two separate methods. There’s creating a descriptive view for which you are using to create the model, and then there is operationalizing the model. On the operational side, it’s data coming from the Thing model being scored against the actual predictive model that’s created. For the descriptive view, the tool of choice ultimately boils down to organizational preference. How you load represented data into the Analytics Server is completely based on the tools with which you work.

 

CAN YOU EXPLAIN IN DETAIL THE STEPS FOR CONNECTING A MACHINE TO THE THINGWORX ANALYTICAL MODEL, INCLUDING HOW TO DEFINE THE DATA COMING FROM THE MACHINE TO CREATE THE MODEL?
Absolutely. I’d encourage you to go check out our new Operationalize an Analytics Model developer guide available on the ThingWorx Developer Portal. In just 30 minutes, you’ll learn how to use Analytics Manager and ThingPredictor to automatically perform analytical calculations.

 

OUR ORGANIZATION SEES FEATURE ENGINEERING AS A KEY PART OF THE DATA ANALYSIS PROCESS. DO THINGWORX ALGORITHMS HANDLE FEATURE ENGINEERING INTERNALLY?
Yes. There’s feature engineering in terms of getting a dataset ready for consumption. The technology ThingWorx provides is being able to automatically sift through the data and use various features to guide the selection of algorithms. Feature enrichment is what’s really powerful about our supervised machine learning capabilities. 

 

HOW DO I INCORPORATE ADVANCED STYLING IN MY UI, LIKE ANIMATIONS AND RESPONSIVE BEHAVIORS?

The standard way of achieving advanced styling in ThingWorx is to leverage the Media Entities, Style and State Definitions. Many widgets, such as the Value Display and the Shape Widget, have out-of-the-box ability to take a State Definition and apply advanced styling for things like severity of risks, etc. Watch our IoT Application Makeover webcast for more information about this topic.

 

DO YOU HAVE ANY RECOMMENDATIONS FOR GUARANTEEING DESIGN CONSISTENCY IN MY IOT APPLICATION?
For non-designers: to keep your design clean and consistent, it is important to properly manage your Style Definitions. Define styles and stick to re-using them; don’t have five different styles for showing model accuracy, for instance. I would recommend creating 5-10 styles for text, and then from there choosing a color scheme for things like buttons, labels, charts, grid headers, and other elements where you need a vibrant color.

Version history
Last update:
‎Apr 03, 2018 03:10 PM
Updated by:
Labels (1)