This thread is for posting solutions to “Just KNIME It!” Challenge 11. Feel free to link your solution from KNIME Hub as well!
Here is the challenge of the week: Just KNIME It! | KNIME
Have an idea for a challenge?? We’d love to hear it! Feel free to write it here.
Quick and dirty solution with a Histogram and Data Explorer to get mean and median ticket resolution time!
Ticket Mean Time and Viz
My submission for Challenge 11 for Histogram and Mean calculation
Bit straight forward … I hope i am not missing any inputs
here is my solution to the weekly challenge: justknimeit - 11 - Raffaello Barri – KNIME Hub
I made computations considering the difference between closing and opening date. This is right, but there should be another measure that excludes holidays, Saturdays and Sundays. I found a way to do it in Python (which I can’t use) and another one in R, but the latter works only with date (not datetimes). I was not able to implement it yet. Maybe I can try to do it in the following days.
This one was easy indeed!
Here is my solution:
For visualizations I used 1 histogram and 3 text output widgets for total ticket count, open ticket count and mean ticket duration time (minutes).
I decided to show the average time spent on a ticket monthly.
Solution attached. Added some basic daily statistics to see if more granularity would explain anything.
REF Challenge 11 Rev 1.knwf (843.9 KB)
here’s my solution
I created a few histograms (overall and split on weekday/weekend) and bar chart summarizing avg. completion time per month of year and time of day
Here is my solution to #justknimeit-11 :
KNIME Hub > gonhaddock > Spaces > Just_KNIME_It > Just KNIME It _ Challenge 011
Task (1) From the histogram display, it can be assumed that we are not dealing with a normal distribution.
Task (2) As in the rest of answers, I’ve tested to calculate the means on different granularities. In the same way, only monthly averages give some feeling of trend…
Moving forward on histogram analysis, displaying the histogram as LN values of time in seconds gives a better scale for the analysis, and a straight forward method to deal with like-outlier values in decimal scale. This chart shows box plots for Solving Time [Seconds] plotted in decimal vs. LN
Representing the histogram as the LN of solving time in seconds, it shows a better scale for analysis… however histogram frequency plot shows a tri-modal distribution.
It is hard to move forward on this based on a single parameter; I’ve decide to assume 4 Ticket families with different Mean value in solving time. I ran a clustering based on a single parameter ‘LN_Solving Time [seconds]’ (!)
Anyhow, these are my Total Mean value [Hours.Decimal] results for the different clusterized ticket families.
my approach for this challenge. Besides calculating the required measures I wanted to play around some more with a couple of visualizations that I always fund useful. Trend lines and month vs month comparison.
Even with the simplest dataset we can achieve very interesting statistics and visualizations !
Whoah! And true that. Amazing work!
We REALLY appreciate how you folks go above and beyond when offered a simple challenge!
Here’s our solution to this week’s challenge!
As you can see, we kept it simple: no assumptions regarding weekends or holidays, and we also did not use metrics that are more robust than the mean given how the data is distributed. Indeed a baseline for the problem — and we are so happy to see how you folks built way more complex solutions!
Way to go, community!
Keep KNIMing and we’ll see you tomorrow!