If you are interested in showing the change over time then the best chart is the good old line chart. I noticed that you are trying to show the average speed by using a variation in the colour but this is hard for people to understand. Position is a far better indicator of value, followed by size and then by colour. I have used a bar chart in the example below as it shows up the colour scheme better.
Secondly, I think doing it to the minute might be a bit too granular. I have tried it to the hour and it gives an interesting pattern to the data. This will depend on what patterns you are trying to identify.
Thirdly, you might want to look into using the Trellis Chart option, on the general tab of any chart. This will take a dimension and use it to loop through, creating a version of the chart for each value. I have used it in the screenshot below with Geographic location used as the extra dimension..
Let me know your thoughts and we can tweak it further.
I agree that the simplest option is the Line chart, however a line chart doesnt give a true representation of the data.
Above is what im trying to replicate, currently I do it with a VBA macro, what it shows is the speed data on the greyscale and when signals are set. (Blue just means no data). The idea is to make it very easy to see when and where there is a queue building and how the signals react to it. Im trying to replicate this in qlikview so i can incorporate the drill down functionality and interrogate a longer time scale over a larger stretch of road.
I've seen this called a Ribbon Chart. Not a native QV chart type which would explain the problems.
I think you can achieve something similar using a pivot table. Create dimensions of Geography and time with an expression of "1". Then drag the time column label across so it becomes horizontal, essentially making a cross table of Geography and Time with each cell value as 1. Then apply your colour scheme to the expression to give the cells the correct colour shading.
You might hit some problems because of the amount of data and that enormous If statement. Perhaps you might be better off using the ColorMix wizard to make your colour scale, or calculate the colours as part of your load script so it doesn't have to be worked out on the fly every time a selection is made in the app. I would probably use the If statement for the blue colour and then if there are no special conditions use a colour value already calculated in the load script.
(as an aside, I've tried this with the app you provided and the performance is pretty poor with the number of columns you would need to show. I'm not sure if this is a data model issue, something to do with the pivot table or possibly a bug. I'll keep playing and post a solution if I find one.)
You could also use a show condition so the chart is only seen once a date or time period has been selected.
Thank you for the advice, doing the calculations in the load script probably makes the most sense. I may have to resort to the custom table though as the data i/we have been working with is a sample, its a fraction of the size that i would need to use.