Hello, I'm Sekine of the digital Business Promotion Office (dejvi).
Last by doing process mining using the log of the intra-mart
Which time zone, it is possible to check (screen transition) access have focused on which screen.
This time, using the transaction data of intra-mart process mining
Will introduce the method to carry out.
Goals in this article
By using the transaction data of IM-BPM and perform the visualization, analysis of the process.
Process mining using transaction data
Data that is used in this article
Data that is used in this article
By using IM-LogicDesigner and job
Will be the stored data by IM-BPM move automatically.
※By setting the exact performance to task for each and every user
Can also be used in the simulation or
And can also be used in the load test
About it, I would like to the article next time.
First of all, Format the transaction data of intra-mart in KNIME.
※This by using the Database Reader
Retrieving data directly from the database of intra-mart.
Order to which you have received the data to some extent reshaped in SQL
KNIME flow itself is the simple.
Visualized by Disco
Again this time by using the Disco, you capture CSV.
As a result ripped, it will be as below.
◆ Result ripped to Disco
◆ The original IM-BPM
Vertical/horizontal has been altered, to the same flow of IM-BPM and.
Analyzed by Disco
By bringing in the Disco, the frequency of the task
You are able to view the task residence condition, performance, and animation.
Can also be analyzed by various ways other is possible by using the Filter function.
● Preferred update frequency and performance visualized task
Left 'frequency' (task frequency), right are displayed 'performance (mean) ' (average Proess Time).
From this map, by looking at line weight and color of the box, you will see a visual load of "officers platform approval" is high.
● Task analysis of time series
Can be observed in the time series peak (at around 11) in the morning, the peak (at around 15) in the afternoon, the change of the convergence (at around 18) and tasks.
● Frequency and performance for each task
"processing status query" Frequency of the high, you can see that are cook task properly.
On the other hand, "officers platform approval" in addition to the height of the frequency, since the processing time is longer, there may be some bottlenecks to this task.
● Frequency (task assignee number) of resource (people) for each, and performance
"mr. Harada" Will be and know processing time of "okada-san" very long.
Specifically, "mr. Harada" large amount of the task, which is in charge of can also load is applied to the "mr. Harada" alone here.
● Analysis was changed cut filter
This by which you want to filter by "officers platform approval", we carried out checks the user is handling this task.
"officers platform approval" Not only 2 name of the "mr. Harada" and "okada-san" in-charge, and processing time is not cook number for the long task is assumed.
※By using a filter function
"in case made via which activity is takes time"
"in case who process is takes time" Analysis such as is also possible.
● How the change process together with the flow of time
By in the animation
Can be watched the gestures of? What time stagnation of which task has occurred.
※In orange of 0 residence
of the task is represented.
※White 0 is in progress task
is in the box.
● Analysis results
As this analysis
・Many task amount of "officers platform approval" as a whole flow.
・"officers platform approval" is a business processing takes time.
・"officers platform approval" is performed in 2 patients.
・Many task amount of "mr. Harada".
That. I understand.
From this result will likely lead to business improvement by doing the following.
・Whether the business practices not impossible.
・Or cannot increase in-charge for the "officers platform approval".
・Segmentation of the task performed, for example, cannot to load reduction by introducing RPA.
Process mining using transaction data together
Approach to capture the last trunsition log
Has been frequency of access as a screen transition, see
This time, you will use the data of intra-mart IM-BPM process mining
By performing, it is possible to perform visualization and analysis of business process could be.
Is a little strange story of process mining from the data of BPM
Because there is no noise data, I think that you image of the visualization and analysis of the process easily.
Osaka October 9, 2019 (water)
Tokyo October 23, 2019 (water)
Osaka is held in The Ritz-Carlton Osaka!
Tokyo is held at The Prince Park Tower!
In digital Business Promotion Office and perform the DX exhibition in succession to the last year in Tokyo.
You came to see the below sentence… Please come to DX exhibition booth!
・Digitizing analog work, the process of manual business, visualization
・Sorting of the task efficient by AI
・Closest to the success with AI discovery of business process
-> EWS2019 special site ←