Thermos flask

PeteHall

Moderator
pwhole said:
Edward Tufte would be laughing his pants off at that graph - it's hilariously not useful.

I disagree there.

The graph shows data at 5 points in time; 0mins, 90mins, 4 hours, 8 hours and 24 hours. Each of those times is clearly shown on the x-axis. You could very easily work back from this graph to the original data, but it presents it in a more visually useful way than simply showing a table of data. I agree that it's not how you would present the data in a scientific paper, but it does tell you something useful, unlike the Fenix performance graphs for example...
 

Speleofish

Active member
I agree with Pete. Provided you read the numbers on the x-axis correctly, it's quite clear. Much better than the standard rubbish graph which shows only a small piece of the y-axis to exaggerate small differences.
 

andrewmcleod

Well-known member
It's not *deliberately* misleading in any way, it's just the wrong way to present the data. It's like missing capital letters from the start of your sentences or something.

In fact there are a wealth of things wrong with that graph :)

1) the tick marks on the x axis are between the labels, rather than on the labels (non-critical, but why?)
2) there are too major grid lines on the y axis (using just a few strong grid lines would make it easier to read)
3) it would be good to have symbols (e.g. crosses) at the data points themselves, as that's what the real data is
4) the resolution is crappy and there are horrible (JPEG?) artifacts; probably a copy-paste rasterizing issue
5) too many of the colours are too similar: lots of yellows and oranges (probably there are just too many lines, but dashes etc. could be added for disambiguation), although you can most disambiguate by order which is good

But they do at least have units on the axes, so that's a plus :)

Generally I would say that using lines to join points like this is a bad thing generally, but particularly when using non-linear axes like this it should be a crime. However, in this case due to the number of lines it does make it a lot easier to follow the values of individual flasks. It probably wouldn't have hurt to put the values at the right points on the axis though...
I suspect though its just that the journalists making this graph didn't know they should use the 'scatter' graph (with best fit lines, probably) rather than the 'line' graph (which is almost never the best choice)...
 

kay

Well-known member
it presents it in a more visually useful way than simply showing a table of data. I agree that it's not how you would present the data in a scientific paper, but it does tell you something useful
It also appears to tell you something that isn't there in the data. It's even more important to get graphs right for a lay audience, who can't be relied on to read the labels on the graph, but will merely look at the overall picture, which in this case appears to show that all the flasks (except for the bottom one) have a steady decline in temperature to 8hrs, and a much steeper decline thereafter.


It's misleading in that they have joined the points with a line, thus implying a a trend and inviting the viewer to see an increase in the rate of drop off after 8 hours. That is as misleading as, for example, using pictograms whose height indicates the variable measured, when the eye sees area and misconstrues something to be 4 times as large when the height is only twice as large. Yes, you can pore over the graph and get back to the original data, but the point remains, the graph is showing a story which isn't there.

using lines to join points like this is a bad thing generally, but particularly when using non-linear axes like this it should be a crime
Spot on!

 

Chocolate fireguard

Active member
I agree with most of the recent comments.
When I was a productive member of society one of my jobs was teaching physics.
The idea that experiental results are just a series of snapshots of what is happening, and that filling in the bits between those snapshots requires an understanding of the physical process was difficult for some to accept. Just joining the points up is a lot easier.
In the case of the flasks, if the heat loss is purely via conduction to the outside air then that will produce an exponential change and a graph with water temperature minus ambient temperature on y and (proper!) time on x would show this.
Although sketching in the curve with only those few points would be challenging!
I like to think that some of my brighter and more devious pupils would have plotted a log temperature versus linear time graph, confirmed that it was a reasonable straight line, invented some extra points along it then worked backwards to increase the number of original results and make it easier to sketch in the exponential fall.
 

yuvals

Member
My 2 Cents: whatever brand you choose,  pre-heating of the flask before purring in the beverage, can make a big difference in how long it keeps the beverage hot...   
 

tomferry

Well-known member
So I have decided to go for a thermos king for her I will advise her to pre heat it before use each  time ,  also I have found her loads of wool kids jumpers in a charity shops so she can make all different covers for it . Thanks everyone  (y)

https://thermos.co.uk/stainless-kingtm-flask-470ml

I was tempted by the Swiss 1 have to say .
 

kay

Well-known member
HardenClimber3 said:
Perhaps the x axis should have been log. Oh wait, that would be non-linear and thus criminal....

I think you?re joking, but ? you?d use a log axis if you felt the underlying process was multiplicative rather than additive, as a way of turning a curve into an easier to work with straight line. And you?d tell people what you?re doing. This article has used a non linear axis to introduce a curve into the line, and hasn?t bothered to tell people. It?s aimed at an audience who will read it fairly quickly, who probably won?t scrutinise the axis, and will go away with a picture of thermoses being OK for 8 hours then declining rapidly. Since this equates (for other reasons) to experience, ie the lunchtime drink is piping hot, but the half cup you save till the end of the day is barely lukewarm, nothing is going to say to them ?wait a minute, that doesn?t make sense?
 
Top