Best 4-Factor Graphic Impactor

Hello,

I'd like to know which chart type is best to visualize 3-4 data sets with some mutual dependency that exerts one cumulative effect on a single outcome. For instance, if you wish to show how Time, Weekday and/or Count may have direct impact on Total Amount of revenue as depicted below.

Thanks in advance!

You might consider trying a radar chart which is good for looking for differences in multiple parameters.

MRice,

Thanks for the advice. Cursory research suggests radar charts work for just 2 variables rather than 4. Or, is there some way to combine all 4 into just 2 data sets? What's the best method for my scenario?

We have always had difficulties visualizing multidimensional data in a 2D space (screen or paper). A lot will probably depend on exactly what you are trying to do, how much is about the data visualization part of the problem and how much is using data visualization as an aid to data analysis. Who your audience is and what you are trying to really show.

My first thought is to maybe step away from trying to visualize 4D data in a single chart. My first inclination is to separate out 3 different dimensions and create different "contour plots" -- plots where I will have the dependent variable (hourly amt?) on the vertical axis, one independent variable on the x axis, and each data series in that chart is a second independent variable. Each chart will ignore one of the independent variables.

Starting with an eye towards using visualization as an aid to data analysis, I started by clearing the filter from the top table, selecting A3:H26, and inserting a chart (for now any chart will do). This gives me a chart with hour/time of day along the x axis, hourly amt along the y axis, and each data series is a different day of the week. I try several different chart types, and I settled on a stacked column chart, which seems to show pretty well that hourly amt is higher in the evening and overnight hours, and lower during the day. A quick Switch Row/Column command gives me another stacked column chart with weekday along the x axis, hourly amt on the y axis, and each series is a different time of day. This chart suggests that hourly amt is higher on Th and Fr than on other days of the week.

To look at the effect of count, I create a scatter chart using the pivot data in rows 36 to 61 (which means I need a technique like this to avoid creating a pivot chart: https://peltiertech.com/regular-char...-pivot-tables/ ). By creating 7 data series from the individual pairs of CNT and Hr Amt$ columns, I can get a scatter chart with CNT on the x axis, Hr Amt on the y axis, and each series is a different day of the week. I see a strong correlation between CNT and Hr Amt, but I don't see much difference between days of the week.

Those were the three charts I could quickly create from the data in your sample file. With more work, I'm sure other charts could be created to see what other trends might be shown. Of course, this only works if you are amenable to using multiple charts rather than one single chart. Some of the details will depend on your audience and exactly what you want to show them. It might give you some ideas to work with.