Skip to main content
Announcements
Qlik Connect 2024! Seize endless possibilities! LEARN MORE
cancel
Showing results for 
Search instead for 
Did you mean: 
datanibbler
Champion
Champion

Best chart_type to display this?

Hi,

I have the requirement to display the distinction between several temporary_employment_agencies with all our temporary workers

=> There are 12 or so different companies we have hired workers from, so I thought that with a barchart, the nr. of segments would become too big, making the whole thing difficult to read.

=> I was thinking that a pie_chart would be best to display the nr._of_workers_from_a_given_company.

<=> However, I also have to use the dimension "day" or "month"

Can anybody tell me what would be the best chart_type to use here?

Thanks a lot!

Best regards,

DataNibbler

5 Replies
kalyandg
Partner - Creator III
Partner - Creator III

hi,

Line Chart is the best type for periodical analysis

Regards,

Kalyan

Anonymous
Not applicable

Nibbler

Hard to say without seeing the real data, but I would suggest a pivot table.

If there is too much data for a bar chart to be meaningful then a pie chart would be even more hopeless.

Stephen Few has written some great books on Visualisation and I would recommend them all.  He has a particular disdain of Pie Charts that is well worth heeding.

Best Regards,     Bill

ngulliver
Partner - Specialist III
Partner - Specialist III

Hi, DataNibbler.

As Kalyan mentioned, line charts are the most suitable for changes over time. However, with 12 or more different companies it may not be very clear. What you could do is go onto the Dimension Limits tab and only show the most important companies (top 5 ?).

Pie charts are not very helpful for large numbers of slices. Stacked bar chart can be an alternative to line charts but, again, may look rather messy with many companies.

The suggestion of a pivot table is certainly good at displaying the information but isn't ideal for showing trends over time.

So really the solution depends on what you want to focus on. If it is time, then a line chart, if it is volumes then a pivot.

Cheers,

Neil

datanibbler
Champion
Champion
Author

Hi,

thanks a lot for your advice!

I would happily read up on visualization - unfortunately it's only partly in my hand: Management has already stated their preferences and going forward they will also do so and, well, at the end of the day it is they who pay my salary, so it's their will that counts... I am not happy with all of it - IMHO, for example, we use too small a fraction of the available screen_space - about 1/3 - to display the actual working data ...

Also, because I am not only designer, but (mainly) developer of the apps (and QlikView_admin, too), I cannot delve too deeply into details. Still, I'd like to. When we have some budget again next year, I'll get some material to read up. I spend a few hrs on the train every day, so there's time.

You're probably right about a pie_chart becoming hard to read - we have employees from approx. 20 different agencies (at the moment).

Still, I think that is the easiest (most visual) way to display that - leaving out the time_dimension and just taking a snapshot from yesterday. - I will give the user some possibility of restricting display to, say, 90% of the total.

Best regards,

DataNibbler

@ Kalyan

Thanks for your answer as well! A line_chart, however, would also become quite hard to read because of the nr. of lines. That's actually a problem in the data, but I have to cope with it somehow...

datanibbler
Champion
Champion
Author

Hi Neil,

I will close this thread - the fact that decisions on visualization and layout don't altogether lie with me makes this somewhat pointless.

Thanks for the answers  anyway!

I agree with you - a linechart or pivottable would probably be best as I cannot restrict and only show the top5 (that would immediately raise questions 😉 - but still I will go with a pie_chart as it is the most "visual", "sexiest" kind of chart to use. It looks a bit confusing with 20 different companies, but I just cannot leave out any data.

Best regards,

DataNibbler