Need ideas on how to plot this large set of data

[DyslexicHobo]

I'm working on monitoring and trending the performance of a memory controller. Without getting into the technical details, I'm trying to find a good way to visualize the data I've collected in a way that will be easy for me to update when I receive new data.

There are about 450 individual sub-module controllers that I'm monitoring about once a month. The value that's being reported each month is a value anywhere from 5 to 30,000 (most are <100, and the ones I'm most interested in visually representing are the ones <100). I have about 200 data points for each of these 450 individual sub-modules, and I just can't figure out how to plot this data in a way that's visually appealing while not leaving out important information.

I'm not really that well-versed in data visualization. I know the basics of Excel but that's about it. I know it's silly to plot all 450 components on one line graph, so I'm hoping there's some plotting method that might apply here.

Let me know if any other info/clarification would help!

 

[jedishrfu]

so you have 450 modules with 200 months worth of data (ie one datapoint per module per month)?

The only thing I can think of is a color chart with x=month and y=module number and each datapoint as a color (blue <100, green >100 to <1000 and say red > 1000 or whatever color scheme you want. Basically you'd have a map of time vs module number horizontal red lines would indicate a module that has consistently had higher values and vertical lines would indicate a group of modules that had higher values for the given month.

Basically if you know what you're wanting to see then design the chart to show it. As an example, maybe you only want to see those modules with high values in the month so just drop the low value ones. Or maybe you want to plot value ranges vs number of modules within the given range.

 

[DyslexicHobo]

This is a great idea! I think I'll use this and just omit all modules with consistently low values.

Thanks for the idea.

 

[jedishrfu]

This is a great idea! I think I'll use this and just omit all modules with consistently low values.

Thanks for the idea.

Great!

I was also thinking if there are any other features of the modules that could be charted such as location vs recorded data values. As an example, temperature modules being mapped to a physical map showing temperature variations from place to place. Another would be if the modules use different versons of software perhaps grouping them that way to compare software versions vs data values...

 

[CellCoree]

One possible approach to plotting this large set of data would be to use a scatter plot. This would allow you to plot all 450 sub-module controllers on one graph while still being able to see the individual data points. You could also use different colors or shapes to represent different ranges of values, such as <100 or >100, to make it easier to identify trends and patterns. Additionally, you could consider using a trend line or regression analysis to show the overall performance trend of the memory controller. Another option could be to use a heat map, where the color intensity represents the value of the data point, to quickly identify areas of high and low performance. Whatever method you choose, it's important to keep in mind the purpose of the visualization and ensure that it effectively communicates the information you want to convey. It may also be helpful to consult with a data visualization expert or do some research on best practices for visualizing large datasets.