Skip to main content
Question

Generate an average graph for each day

  • May 26, 2026
  • 3 replies
  • 45 views
J
Explorer

I want to create a graph which represents a typical day for my process, as there is a pattern in the values during each day. So I would like to average the values at a specific time during the day, for the last 30 days or the last year, and create this average value for each timepoint during the day. How would I go about this? The only way I could find to average values is over a single continuous time window, which is not what I want unless I could average several of these time windows and make each window small enough to result in a clear typical day graph (1 minute or 10 minutes for example).

 

    3 replies

    F
    Employee
    Forum|alt.badge.img
    • fvandael
    • Employee
    • May 27, 2026

    Hi ​@Jille,

     

    The approach that works well here is combining a Value-Based Search on the TM_hour time tag with Event Analytics Scatter Plot.

     

    Step 1 — Set up the search

    Open a Value-Based Search, set your context bar to the history window you want (-30d or -1y), then add TM_hour_Europe_Brussels (or the timezone variant for your site) with condition Constant and a minimum duration of 2 minutes. This produces one event per contiguous clock hour — 24 events per day, ~720 for 30 days, ~8,760 for a full year.

     

    Step 2 — Add two calculations

    • End value of TM_hour_Europe_Brussels — gives the integer hour (0–23) at the close of each event.
    • Average of the process tag you want to profile or any other calculation available.

     

    Step 3 — Open Event Analytics → Scatter Plot

    Switch to the Scatter Plot tab and set:

    • X-axis → End value (TM_hour_Europe_Brussels)
    • Y-axis → Average [your process tag]

    Each dot represents one hour-block from one day. The 24 vertical columns show the distribution of your tag at each clock hour — the typical day shape is immediately visible, and the spread of each column tells you where variability is highest.

    This workflow is also described in the community here (steps 1–4 of the multivariate insights post map onto exactly this approach):

    The TM_hour_* and TM_day_* family of time tags are also described in:

    Kind regards

    Frederik Vandael

      yracette
      Pioneer
      Forum|alt.badge.img+2
      • yracette
      • May 27, 2026

      There is a lot that cane be done by leveraging the TM_hour* , TM_day*, TM_month*, TM_year* demo tags for sure, and event analytics offers nice visuals around this. I wish we could include an event analytics set of visuals on dashboards directly much like we now can with the statistics tables.

      Yves Racette
      J
      Explorer
      • Jille
      • Author
      • Explorer
      • May 28, 2026

      Hi ​@Jille,

       

      The approach that works well here is combining a Value-Based Search on the TM_hour time tag with Event Analytics Scatter Plot.

       

      Step 1 — Set up the search

      Open a Value-Based Search, set your context bar to the history window you want (-30d or -1y), then add TM_hour_Europe_Brussels (or the timezone variant for your site) with condition Constant and a minimum duration of 2 minutes. This produces one event per contiguous clock hour — 24 events per day, ~720 for 30 days, ~8,760 for a full year.

       

      Step 2 — Add two calculations

      • End value of TM_hour_Europe_Brussels — gives the integer hour (0–23) at the close of each event.
      • Average of the process tag you want to profile or any other calculation available.

       

      Step 3 — Open Event Analytics → Scatter Plot

      Switch to the Scatter Plot tab and set:

      • X-axis → End value (TM_hour_Europe_Brussels)
      • Y-axis → Average [your process tag]

      Each dot represents one hour-block from one day. The 24 vertical columns show the distribution of your tag at each clock hour — the typical day shape is immediately visible, and the spread of each column tells you where variability is highest.

      This workflow is also described in the community here (steps 1–4 of the multivariate insights post map onto exactly this approach):

      The TM_hour_* and TM_day_* family of time tags are also described in:

      Kind regards

      Frederik Vandael

      How would I then go about averaging the values at each hour so I'm left with a single dot for each hour?

      Need help?

      Can't find the answer to your question? Create a new topic or file a support ticket.

      New to TrendMiner?

      Discover our Industrial Analytics platform at www.trendminer.com.

      Terms & ConditionsAccessibility statement