Which type of graph would be best to use for...(Please help, I always confuse graphs)

Question

Which type of graph would be best to use for displaying each variable (combined tackles and position) for the 2014 NFL tackle stats graphically and what evidence supports your choice? ( You may choose from Time Series, Bar Chart, Segmented Bar Chart, Circle Graph, Dot Plot, Histogram, or Ogive.)

Answer 1

Determining graph type is mostly dependent on what kind of data you have. In this case, dot plot, histogram, and ogive are out immediately because our data is categorial. It's also definitely not a time series. So we're left with Bar, segmented bar, or circle.

I don't think I have enough information here to choose the right graph, honestly. My assumption is that you have combined tackles by position for a set of teams? or something like that? In that case, a segmented bar chart makes the most sense- you'll want to see how tall each bar is for each team and what portion of that team's tackles came from what position. If you don't have teams and only have tackles by position, then a regular bar chart might be best. If you're trying to see what position makes up most of the tackles, a circle chart might be okay (but a bar chart is probably better in that case too).

Let me know if you need any help here. I'm happy to explain further if you need it!

Answer 2

So maybe let's think some rules of thumb to remember the "best" way to represent different types of data (which can be confusing sometimes). Using the correct graphs as a data scientist or statistician can strengthen the persuasiveness of an argument or help your audience understand a concept better!

Let's take a look at the NFL stats, combined positions and tackles.

You would use a bar chart to plot:

Categorical data, like number of tackles per position

Linebacker - XXXXX

Quarterback - XX

Tight End. - XXXXXX

You could also use a histogram in this case, since you are only interested in the distribution of tackles (one variable). Each "bin" would be a position the NFL player plays, and you would stack one block in each bin for each tackle.

You could use a segmented bar chart if you wanted to add a second category like "missed" versus "successful" tackles. You could use one color to describe category "missed" and another for "successful".

Linebacker - XXXXXOOOO

Quarterback - XXOOO

Tight End. - XXXXXXOO

Dot plots are very similar to histograms in that they both look at count data. However, histograms look at the number of times something occurs (tackle) per category (position). Dot plots consider each occurrence -including ALL values from the data set, with one dot for each occurrence of an observed value from the set. These dot plots are best for small datasets - probably not the number of tackles in the NFL :)

You would use a time series to plot if:

You wanted to know the number of tackles per position per game. So your x axis could be Game 1, Game 2, Game 3, Game 4,...,Game 10 and your y would be number of tackles at that game. You could use different colored lines or different shapes for the points for each player.

Linebacker (Game 1, 2), (Game 2, 0), (Game 3, 5),...,(Game 10, 1)

Quarterback (Game 1, 1), (Game 2, 3), (Game 3, 1),...,(Game 10, 5)

Tight End (Game 1, 5), (Game 2, 6), (Game 3, 8),...,(Game 10, 3)

You could use a circle graph if you wanted to show the percentage of tackles each position contributes to total tackles (% contribution of each category to the whole).

50% Linebacker

10% Quarterbacks

40% Tight End

An Ogive plot would help you find the median number of tackles per position. You plot the cumulative number of tackles "more than 2", "more than 4", "more than 6"... and "less than 2", "less than 4",... etc to gain information about the frequency of the occurrences. I don't think it would be appropriate in answering this question - here's a good reference for when to use these (https://byjus.com/maths/ogive/).

** The point **

Without more context on this question, I would say a histogram or a bar chart would be most appropriate for representing the data. Hope this helps!