Business Context
An industrial manufacturing plant monitors engine temperatures via IoT sensors. The data is continuous, but engineers need to categorize temperatures into 5 distinct 'bins' (ranges) and perform a Chi-Square test to check if the temperature range is statistically independent from the machine's 'Status' (Operational vs. Error).
Data Preparation
Simulation of 5000 sensor readings with continuous Temperature and Status.
Copied!
| 1 | |
| 2 | DATA casuser.sensors; |
| 3 | call streaminit(999); |
| 4 | DO i=1 to 5000; |
| 5 | Temperature = rand('NORMAL', 80, 15); |
| 6 | IF Temperature > 100 THEN STATUS='Error'; |
| 7 | ELSE STATUS='Operational'; |
| 8 | OUTPUT; |
| 9 | END; |
| 10 | |
| 11 | RUN; |
| 12 |
Étapes de réalisation
1
Run crossTab with automatic binning (5 bins) on Temperature and Chi-Square test.
Copied!
| 1 | |
| 2 | PROC CAS; |
| 3 | SIMPLE.crossTab / TABLE={name='sensors', caslib='casuser'} row='Temperature' col='Status' rowNBins=5 niceBinning=true chiSq=true; |
| 4 | |
| 5 | RUN; |
| 6 |
Expected Result
The output should display the 'Temperature' variable grouped into 5 intervals. It must also include the 'ChiSq' table containing the p-value, confirming the strong correlation between high temperature bins and 'Error' status.