Our World Statistics Day conversations have been a great reminder of how much statistics can inform our lives. Do you have an example of how statistics has made a difference in your life? Share your story with the Community!
Choose Language Hide Translation Bar
Making sense of sensor data with JMP

Scientists and engineers who work with high-density sensors face data problems that can make getting insights from data more challenging. Whether you’re trying to make sense of information from industrial devices living on the Internet of Things or monitoring health and fitness parameters, JMP provides an ideal sandbox for sifting through noisy data to identify important patterns.


  • Database Query Builders. When your sensor data lives in a structured SAS or ODBC-accessed database, you can quickly create queries that join multiple tables, apply filters, and preview the results so you always get the perfect subset to explore. Queries can be saved, shared, and even updated directly from your imported table. You can move more quickly from raw data to insight by adding summary columns and graphics scripts to run automatically post-query.
  • JMP Query Builder. If you don’t have direct access to your database, you can still take advantage of the query, filtering, summarization and sorting tools afforded by the Query Builder by importing a set of flat files into JMP tables. After you create a query once, it can be saved as a script and rerun automatically. Just pull the most recent data, rejoin, and move forward with your analysis.
  • New Formula Column. Often, sensor data is collected and stored simply with a time stamp, sensor name and sensor value to conserve storage space. But pairing raw data with summarized values can help you answer broader questions, clean up artifacts and discover interesting trends. Right-click any data column to add a new formula column, choosing from Date/Time transformations and summaries like Col Moving Average, useful for cleaning up sensor data with high-frequency noise. The formula shortcut menu offers quick access to the Group By functionality that identifies a column of categories (for example, Day) by which to calculate new measures.
  • Virtual Join. When working with high-resolution sensor measurements, you may not be able to spare the memory required to merge summary values directly into your raw data table. Use virtual join to link values stored in separate tables, and use them together in analysis and graphing platforms. You can link and unlink data sources on demand, which makes it easier to update disparate data sources independently.
  • Process Screening. Graphing many sensor measures can be tedious when generating control charts is a one-at-a-time process. And once you have your results, sorting through charts manually to find interesting or suspect readings is both painful and error-prone. The Process Screening platform does a lot of the heavy lifting up front, consolidating control chart statistics and recent shifts into a single table that you can sort interactively to identify important process variables. This helps you focus your efforts and makes is easy to drill down to critical control charts first.
  • Modeling and Cross-Validation. At some point, you need to build a model to separate signal from noise. JMP includes a full suite of linear and nonlinear modeling tools that build useful models on their own but also link into cross-validation tools, crucial when evaluating observational data like those from sensor streams.
  • Graph Builder. When exploring your sensor data, you need to quickly create and customize a variety of multi-element graphs to identify patterns and spot potential artifacts. The JMP Graph Builder platform provides an extensive palette of graph types and an intuitive drag-and-drop interface. Plus, you can save and reproduce any graph you create when you retrieve new data.
  • Selection Filters. Once you have graphed your sensor data, you may want to consider the impact of changes over time, location or another variable. JMP provides traditional list-based filters for exploring custom data slices, but you can also use a summary graph as a filter for one or more graphs. Simply select elements in your filter graph to refocus your report on a category or time frame of interest.
  • Dashboards. Dashboards help you show patterns in your data in a concise and consistent report and assist in communicating key findings to decision makers. With JMP, you can create a new dashboard from a template and drag-and-drop tables and reports onto the canvas. You can save and refresh your dashboard when new data is available or export it as an interactive HTML web report for colleagues who don’t have JMP.
  • Sharing With Scripts and Add-Ins. When you construct a workflow that gives you the results you need, you can easily share your detailed steps with other JMP users. The software automatically generates scripts to reproduce key data manipulation steps and custom graphs. As you become more experienced, you may want to develop custom menus and add-ins to walk colleagues through more complex workflows.

Production systems are really good at telling you what is going on right now, and applying systems of control to run business rules, prevent dangerous conditions and operate systems within predefined spec limits. But a world of insight also lives in your historical data, and JMP can serve as a no-risk sandbox for finding patterns and trends in that data. You’ve gone through the trouble of meticulously collecting and storing your sensor data – now make better use of it with JMP.

Article Labels

    There are no labels assigned to this post.