Since the Falkonry Service was released, numerous users across several industries have used it to understand their #IoT data and recognize conditions. Their objectives are to prevent undesirable outcomes such as failures and low performance/yields. These industries range from power generation and transportation to agriculture, and building management to automotive and steel production. Our users are mostly non-data scientists and seek a solution that they can create and manage. At times, data scientists prefer Falkonry since it can do one thing really well – provide visibility of conditions without much effort.
Falkonry Service–now it’s easier
Taking into account all of the feedback we received from our users, we have redesigned the Falkonry Service to be simpler and yet more powerful. You can more easily work with many things, assessments, and conditions. This new User Interface has been rolled out to all users. Here’s an overview of this new design.
By distilling the raw information down to a one-dimensional timeline, Falkonry simplifies understanding of high speed, multivariate IoT data. The main design improvement in the updated UI is that you can understand Falkonry’s findings, at a macro level, instantly and drill down in terms of time and things easily. As you can see in the design, there’s the ability to compare the performance of two or more things side by side. Simultaneous visualization of conditions across things means that conditions can be interpreted comparatively.
Secondly, both live monitoring and learning views are designed to be analogous, which makes it easier to understand both. The seamless transition from learning to monitoring also means that users can drill-in from high level summary of conditions across an entire fleet, to a single individual thing for a particular period in time where a certain condition arises.
One thing that did not change is that it is just as easy to learn new predictive models without knowing data science or programming. Users can still easily annotate Falkonry’s findings and confirm the conditions visually, as well as in bulk through files.
Next, menus have been simplified so that our users can easily find the pipelines they work with to manage their account. The menu is always to the left and can easily be expanded or collapsed.
Finally, Falkonry’s machine learning continues to improve and find more interesting behavior on its own. As many of our users have told us, Falkonry’s unsupervised machine learning is the best thing to happen to IoT, and we have been working to make it even better. For example, Falkonry figures out from the data what patterns repeat and at what interval so that any periodicity is automatically spotted without the need to give hints or tweak anything.
We continue to make this intuitive experience of IoT data analytics and visualization even more powerful and usable. Even as our technology improves and the Falkonry Service is upgraded, you can be assured that your data and prediction models will be just a click away.
Falkonry’s advanced analytics and simple visualization means that computers are left to do what they are good at – computation – and people can control what the computations do without having to perform detailed debugging or adjustments. There should be no reason for people to pore over hundreds of plots constantly looking for patterns when such work can be efficiently and effectively performed by computers using Falkonry Condition Prediction Service.
This post was originally written February 6, 2016.