Store Time Series of Metrics to Table

 Posted on 2022-07-30  |  3 minutes  |  570 words  |  Kazuhiko Takata

At last post Schedule Notebook to Run Data Jobs I showed you how to schedule Data Job using Machine Learning Workbench. Of course, not only Data Job but any kind of tasks can be scheduled by Python script. Today I would like to schedule backup some kind of data to table by Machine Learning Workbench. Store time series of metrics Imagine I am running Celonis EMS for months and I already established dash board Aalysis to observe some KPIs (metrics). [Read More]
Machine Learning  Data Integration  Process Analytics  Studio  Analysis 

Validate Data Model by Studio Analysis

 Posted on 2022-04-23  |  3 minutes  |  549 words  |  Kazuhiko Takata

At last post Construct My First Data Model, I created Data Model and load data to it. Normally initial load is not perfect, so I should check data in Data Model. Today I would like to share how to validate my Data Model using Analysis. By the way, Celonis EMS main function is now Studio (and App for viewer) and Analysis is also part of Studio, so today I will create Studio instead of Process Analytics to create Analysis. [Read More]
Data Integration  Transformation  Data Model  Studio  Analysis  Variant Explorer  Case Explorer 

Handle Day based Activity as Milestone

 Posted on 2022-03-26  |  3 minutes  |  445 words  |  Kazuhiko Takata

I already created four activities until last post Unite SQL statements by CASE Expression, those are fillfilled requirement of event log. Going back to Consider Case ID before Starting Transformation, case ID is the biggest requirement. Also is is not so big as case ID, but event time is important too. In process mining, event time should be year, month, day plus hour, minute, second (YYYY-MM-DD HH:MI:SS in Vertica format). I guess event time in process mining referred to that is recorded automatically by system responding to something action. [Read More]
Data Integration  Transformation  Process Analytics  Studio  Analysis 

Copy Previous Value to Blank Period by RUNNING_SUM and RANGE_APPEND functions

 Posted on 2021-10-30  |  2 minutes  |  380 words  |  Kazuhiko Takata

At Investigate Workload Trend of Cropped Subprocess I showed trend of activity count, and at that time I used RANGE_APPEND to fill zero count in trend graph. Today I would like to use different aggregation RUNNING_SUM and fill value to blank period. Imagine you would like to check weekly trend of credit amount regarding some customer. Credit amount is increased by the amount of net value when ‘Receive Order’ happened, and it is decreased when ‘Clear Invoice’ happened. [Read More]
PQL  Process Analytics  Studio  Analysis 

Create Key Column of Activity Table

 Posted on 2021-10-23  |  3 minutes  |  460 words  |  Kazuhiko Takata

Celonis Data Model always require unique key in case tabel (case key) to group activities belong to each case. How about activity table ? Activity table do not have explicit key column, instead combination of case key, activity name, timestamp, and sorting number are similar to activity key (those four columns are configured in Data Model). This week I was asked to create activity table key column due to duplication check purpose. [Read More]
PQL  Process Analytics  Studio  Analysis  Data Model 

Investigate Workload Trend of Cropped Subprocess

 Posted on 2021-10-16  |  2 minutes  |  383 words  |  Kazuhiko Takata

Celonis Activity record is good to investigate user workload, so you may already implement analysis to do this. Today I would like to crop activities to minimum subprocess then investigate workload against subprocess. Today’s output image is below. I set filter button of company code, also set three buttons to point out activities to crop subprocess pass through these activities. Right side Process Explorer is to check subprocess, in this case it is starting from ‘Approve Credit Check’ until ‘Cancel Order’ via ‘Deny Credit Check’. [Read More]
PQL  Process Analytics  Studio  Analysis  Process Explorer 

Verify calculation result in OLAP table then convert to visual component

 Posted on 2021-10-09  |  3 minutes  |  435 words  |  Kazuhiko Takata

Today is 24th post of this blog series and I am suprised that I can continue to post blog every week. I got some reply from who loves Celonis in the world and it is my fun to continue posting. Celonis is releasing new functionarity year by year and I am interested in catching up them and thinking how these functions help my clients. After listening to next week’s Celonis World Tour Webiner in Tokyo, I will get more inspiration to write blog. [Read More]
Process Analytics  Studio  Analysis 

Group similar cases by Clustering

 Posted on 2021-10-02  |  2 minutes  |  397 words  |  Kazuhiko Takata

Last week I posted Convert Quantitative value to Categorical one by Quantile Function to create categorical dimension. Today I will create categorical dimension via different way, clustering. Clustering is one of the unsupervised learning method, to automatically group cases by their attributes. Celonis PQL has clustering funciton KMEANS, so you are ready to use clustering. Today I will use O2C process and would like to group customers by (1) their lead time from Invoice send to Clear Invoice and (2) net value. [Read More]
PQL  Process Analytics  Studio  Analysis  Pull Up Function 

Convert Quantitative value to Categorical one by Quantile Function

 Posted on 2021-09-25  |  3 minutes  |  429 words  |  Kazuhiko Takata

Last week I posted Integrate Button Dropdown Entries to one Formula and integrated multiple dimensions. At that time I used string column that enable you to categorize each case, that is called categorical variable. In contrast quantitative variable such as order quantity, net value is not normally possible to use for dimension. If you would like to use these column as dimension, you need to convert its value to categorical value. [Read More]
PQL  Process Analytics  Studio  Analysis  Pull Up Function 

Integrate Button Dropdown Entries to one Formula

 Posted on 2021-09-18  |  4 minutes  |  660 words  |  Kazuhiko Takata

I usually use two types of analysis components, time scale graph and drilldown table for my development projects. These compnents makes it possible to discover root cause of target KPIs by changing time scale or drilldown dimension. In the end, these components and attaching button dropdown were maintained many times, but I was annoyed to set variable value (especially long PQL) to each button entry (it was also cause of defects against my analysis sheet). [Read More]
PQL  Process Analytics  Studio  Analysis