The Student Room Group

Correct type of figure (graph/chart) to use?

I am writing my dissertation and not sure on the most appropriate figure (meaning a graph or chart) to present my data in the most clearest and effective way possible.

If you are good at choosing the right figure to use could you kindly comment please and I will PM you? I don't think what I am doing is 'naughty' but just to be on the safe side I don't want to post my dissertation title on here to the public then get caught for plagiarism (even though chances are slim).

Many thanks
Could you share what type of data it is without giving away what your work is? The great thing about TSR is students can collaborate to give you the best answer :yes:
If in doubt, refer to this:

http://i.kinja-img.com/gawker-media/image/upload/s--mbk6SBVQ--/17mah11legndjjpg.jpg
Original post by Potally_Tissed
If in doubt, refer to this:



OOoooh that's lovely

(although pie charts are never justified :indiff: )
Original post by PQ
OOoooh that's lovely

(although pie charts are never justified :indiff: )


For a simple x vs y they're ok. Should never be used to show more than two things though, otherwise you end up needing to label the segments with their values so that people can interpret it, which rather defeats the point of using a chart.
Original post by Neostigmine
Could you share what type of data it is without giving away what your work is? The great thing about TSR is students can collaborate to give you the best answer :yes:


Categorical data - A positive, negative and neutral category.

And comparing this against more categorical data again for another part. It's quite hard to explain without giving it away
Original post by dylantombides
Categorical data - A positive, negative and neutral category.

And comparing this against more categorical data again for another part. It's quite hard to explain without giving it away


So just take any key metrics and refer to them here as "muffins" or something :tongue:
Original post by Potally_Tissed
So just take any key metrics and refer to them here as "muffins" or something :tongue:


Ok it isn't that easy but i'll try:
I'll call them cakes instead of muffins.

So there are 2 types of cakes and they both will be put into positive/negative/neutral.

There will be subcategories which are a branch off from these cakes. The sub categories will also be classified as pos/neg/neutral. Think of it as slices of the cake so if many of these sub categories are positive, then that type of cake is likely to be positive. But say the subcategories for the other cake seem to be negative, the cake itself is likely to be negative.

That's the best I can do. It may be best if I just actually explain the real thing in a PM.

Anyway if you can actually analyse my made up 'cake' scenario, I was thinking thinking of using 2 figures:

Figure 1 - Using a tree diagram to explain the main thing (i.e the cake), then the branches off for the sub categories, and possibly even another branch off for a third group of categories. The benefit of this is it'll explain how the thing will be analysed from 1 section into the next and possibly a 3rd time (and how each part links to the next). The problem though is it won't display the positive/negative/neutral aspects in a clear way.

Figure 2 - Bar chart like the one I found on google images below. Going across I'll have the categories positive/negative/neutral and going upwards i'll have the percentage of positive/negative/neutral things (i.e "cupcakes"). I think it may clearly present how many are pos/neg/neutr clearer than the tree diagram but having both together would have its benefits.
http://www.lofoya.com/_images/questions/Bar-Chart-lq1.png


EDIT:Another way of potentially doing it would be to create a scale from 0-10. With 0 being negative, 5 being neutral and 10 being positive. Obviously if it falls between those numbers it is weighted slightly more that way (eg 7 would be quite positive but not as positive as a 9 which is extremely positive). Then I will go through all the "cupcakes" and give them a number based on this scale. When all the cupcakes are given a number, I can work out the mean average (add them up and divide by the number of cupcakes). This will give me a number which will be the overall number and will be given to the actual "cake" itself. If I did this though, not sure if I would need to present this pictorially at all.


---

As I said may be easiest if I could explain in a PM
(edited 8 years ago)
If you've got category - subcategory - sub subcategory, that's one too many to display in the same chart unless it's a seriously weird chart. Unless your chart is interactive or you're going to have lots of charts, that detailed level would be best as just a table.

Still not entirely clear on it, but possibly stacked bar charts might work. Ie. A bar for cake 1 and.a bar for cake 2, divided into three segments for positive/neutral/negative. Likewise with the subcategories.

It's not so much your description that's the issue as not being able to see the data itself :tongue: What the data refers to doesn't really matter, but it's hard to suggest anything without seeing the data you're trying to present.

Quick Reply

Related discussions

Latest

Trending

Trending

Articles for you