Utilizing the Custom Floorplan Heat Maps extension in conjunction with Environmental Monitoring

[#1]

In this post, I want to demonstrate how users can utilize a JMP extension to produce helpful visualizations of environmental monitoring data. The following image was generated using JMP Tech Support Engineer Jasean Rasnake’s JMP extension, Custom Floorplan Heat Maps. Using this extension, users can construct heat maps for values measured at multiple locations in a space. An example using a simulated data set provided with the extension can be seen in the following heat map:

It's possible to use this extension in conjunction with results from Environmental Monitoring in the Process Screening platform in JMP to produce similar heat maps, where the processes are color-coded by their alarm status: No Alarm, Alert, and Action. Such visualizations can help users identify where sources of contamination might be originating, providing insight into how contamination spreads between sensors or between different clean rooms.

To produce these heat maps, the JMP extension needs to be installed. For this post, I’m assuming that a user has a floorplan image with process/sensor locations already mapped and just needs to create a floorplan graph. The process to create floorplan maps is detailed in the extension’s JMP Marketplace page.

Creating an environmental monitoring heatmap

1. Use the provided EM Days.jmp data. This simulated environmental monitoring data is provided by the extension. Humidity and Temperature are continuous, and we want to look at a count data case. From Rows, select Data Filter.
2.  Select measurement, making sure the Include option is checked. Then select all measurement sensor types except humidity and temperature.
3. Open the process screening launch dialog from Quality and Process in the Analyze drop-down menu.
4. Next, choose Data as the process variable, Measurement and Room as grouping variables, and Timestamp as the time variable. Next, specify Count for Control Chart Type. Additionally, add symmetric lower quantile probabilities of 0.0015 for Action and 0.025 for Alert under Nonnormal Options.
5. Launch Process Screening. From the red triangle menu in process screening report, select Save Details Table.
6. In this data table, add an additional column called Alarm Status. To do this, go to Cols dropdown menu and select New Columns.
7. Go to Column Properties and select Formula, then Edit Formula. Add the following formula to this column:
   

   If( :Alert Lower Limit <= :Count <= :Alert Upper Limit,
       "No Alarm",
       If( :Alert Upper Limit <= :Count < :Action Upper Limit | (:Alert Lower Limit > :Count) > :Action Lower Limit,
           "Alert",
           "Action"
       )
   )
   

   Note: If you only want to consider upper limits, you can remove the conditional logic involving :Alert Lower limit and :Action Lower Limit. Then the formula would be as follows:
   

   If(Count <= :Alert Upper Limit,
       "No Alarm",
       If( :Alert Upper Limit <= :Count < :Action Upper Limit,
           "Alert",
           "Action"
       )
   )
   

8. Set Property for Value Colors (also accessible under Column Properties drop-down): Select red for Action, yellow for Alert, and green for No Alarm. Alternatively, any other three-color gradient also works. Choose whichever combination you think best suits your scenario.
9. From this data table, select Floorplan Graph from the options under the Add-Ins dropdown menu. Choose Room for location name, Measurement for metric name (this variable denotes the room location and the type of sensor), Alarm Status for measured value, and Timestamp for time. Choose the Lab Floor.jpg file as the image to be used by the extension.
10. Click OK.
11. The resulting Graph Builder report shows you the alarm status for each of the sensors for each time point in the data set. Pictured below is the day 04/24/2024 in the data set.

Environmental Monitoring Floorplan Maps.zip

[hogi]

useful improvement for a future version: Embed Shape Files 
-> will enable/facilitate the sharing of such maps:
just embed the map files into the main data table and send it to a colleague ...
 @martindemel