Histogram in Numbers

Could not agree more.

In all due respect to the experts who have responded to this question. This should be automated. Data distribution is an abstract concept, yet easily and instantly rendered tangible and understandable with software - sadly not with numbers. Alas, the beautiful spreadsheet for the masses, is clearly beautiful, but just a variant of Excel.

If you have ever successfully taught statistics to anybody, you will understand that the shenanigans proposed by the experts in this post are of no use to anybody that could implement them, but are excellent at convincing generations of students that they can't do stats.

The leadership for numbers seems more focused on attractive graphical output rather than promoting numerical literacy. At that they 'excel' (sorry). This is a laudable, but not overly empowering goal.

If you are looking to graphically manipulate data for exploratory purposes - look no further than JMP (by the team that brought us SAS). If you are looking for modern, creative tools to capture, store, and manipulate all sorts of data - try out AirTable at www.airtable.com

And your technical question is... ?

This is a user-to-user technical forum where participants attempt to find answers to questions about Numbers as it is, not as we wish it to be. Previous replies have shown how to create a histogram in Numbers, using what been studied in statistics lessons, rather than the 'black box" of a HIST function.

DG4444 wrote:
I would love to have this automated rather than having to manipulate the data prior to graphing it.

It would be good to be able to just select the graph, select the range and the number of bins and then let numbers do the rest.

This would be a great addition.

That works nicely.

When I use the Reorganize Panel to filter my original data table (by a column I'm not counting), my histogram doesn't change. I'm assuming that the Reorganize Panel only filters what is displayed, and so the functions in my secont spreadsheet aren't affected. Is there a way to dynamically filter the data and have the histogram change, or will I have to make separate histograms for different situations?

As an example, suppose I am tracking the borrowing time for different library books. My data table has two columns, one for the genre and one for the number of days borrowed. I used the method above to make a histogram and it worked nicely, but when I filtered the data table ti display only the Mystery books, the histogram didn't change. What I wanted was a histogram showing only the borrowing times for mysteries.

I think it may be possible to modify the formula to only count the frequencies of entries that match a criteria, but if so I can't figure it out. Am I hunting in vain?

Nope, it can be done, Robert.

The problem, I believe, is that when you filtered your data, you only filtered the DISPLAY of the data, not what data was still in the table. So, FREQUENCY is still including them.

My solution (which includes a little more than you wanted, but it's how I roll) is:

First create a simple 1-column table listing your Genres.

If your books table isn't already populated with hundreds of books, then I'd use an OFFSET function on your genres table to pull in the genre in your books table. This will prevent typos and will be more dynamic should you need to add genres later. That's been explained in other discussions, so I'll leave that alone for now. You might be perfectly happy using a non-dynamic drop-down list.

Now create a small table to house the genre you're wanting to get frequencies for. Again, either use OFFSET against your genres table, or create a drop-down list.

Now add a column to your books table. See my example below:

Run FREQUENCY off of that column and it will change as you change genres in your "Choose Genre Here" table. If you want to include more than one genre simultaneously, that can be done but will require a different means to execute, using checkboxes and a more complex IF function to replace the one above.

Vince