Danny King (kxrs26): Health Informatics Visualisation
Danny King (kxrs26): Health Informatics Visualisation
Page 1 of 2
13/04/2010
Danny King (kxrs26): Health Informatics Visualisation Introduction The software was created using Java and OpenGL (JOGL bindings). I enjoyed this assignment; it was interesting to apply the theory from lectures to a real problem and using OpenGL taught me a lot.
How to use the software To run the software run the DannyKingVisualisation.jar file, ensuring it is in the same folder as the gluegen-rt.dll, jogl.dll, jogl_awt.dll and jogl_cg.dll files. You will need a screen resolution large enough to display the largest window, which is 800 x 700px. You will be presented with the main screen, from which you can choose to view the time series data or the 2007 snapshot data. Rather than list the features you can interact with as bullet points I will label screenshots for clarity.
The time series window Interactive line graph showing the rising level of obesity in England.
Toggle emphasis of the maximum and minimum values for the 3 data sets.
Toggle data display on or off for total population, males or females
Watch the data evolve over the 11year period. One second represents a year’s data addition. You may interact with any of the other features whilst the animation is playing
Hover over any of the data points on a line to see its value appear next to it
The snapshot window Bar visualisation showing the relationship of long-term illness and waistcircumference
Select two of the data sets using the drop-down menus and click ‘compare’ for easier, side-by-side comparison & analysis
Danny King (undergraduate of Computer Science at Durham University) [email protected] www.dannyking.eu
For even easier comparison the difference between the compared data sets is calculated and displayed (green for a growth, red for a shrinkage)
Danny King (kxrs26): Health Informatics Visualisation 13/04/2010
Page 2 of 2
Thought that has gone into the time series data Edward Tufte’s principles of graphic excellence & graphical integrity
316 − 39 Graphical Integrity 39 𝐿𝑖𝑒 𝑓𝑎𝑐𝑡𝑜𝑟 = = 0.963 7988 − 954 Care has been taken not to distort the message of the data; the axes 954 are uniformly incremented and are not misleading in their starting Where 316 and 39 are the number of positions. Points on the graph are plotted at regular intervals. pixels above the x axis of the highest The lie factor is within Tufte’s acceptable range of 0.95 – 1.05 and lowest point of the ‘total’ line Induce the viewer to think about substance, not methodology, design, etc. respectively and 7988 and 954 are No unnecessary graphics, colours, effects or ‘ink’ were used and the the values of those two points from the dataset provided. visualisation has been kept as simple as possible. A line graph was used as most people are accustomed to them, enabling people to focus more easily on the data rather than the design. Encourage the eye to compare different pieces of data Toggling any of the total, male or female data sets on or off allows easy comparison between them. The ability to make the minimum and maximum points to stand out allows for quick comparison & analysis. The time evolution animation allows the user to see data changes over time, allowing easy comparison of how time affects the data. Reveal the data at several levels of detail The ability to toggle the 3 data sets on and off and also to view the data grow over time (evolution animation) allows for different levels of data to be interrogated at different times. Hovering over a point on the graph displays its value next to it, allowing for extra detail. Serve a reasonably clear purpose There is introductory text in the main window which details the purpose of the visualisation. The time series window has legends and labels to provide context.
Thought that has gone into the 2007 detail data Edward Tufte’s principles of graphic excellence & graphical integrity
95 − 61 61 𝐿𝑖𝑒 𝑓𝑎𝑐𝑡𝑜𝑟 = = 0.965 33.5 − 22.5 22.5
Graphical Integrity Where 95 and 61 are the pixel Again, care has been taken not to distort the message of the data; the widths of the female, raised waist widths of each rectangle are all equal. Any percentage value would circumference values for limiting have the same width in any rectangle. and non-limiting longstanding illness The lie factor is within Tufte’s acceptable range of 0.95 – 1.05 respectively (chosen for easier Induce the viewer to think about substance, not methodology, design, etc. calculation) and 33.5 and 22.5 are Again, No unnecessary graphics, colours or effects were used and it is the values of those two pieces of data from the dataset provided. as simple and compact as possible. The user’s eye is drawn immediately to the data which is predominant and in the middle. All colours have meaning. Originally the female data was in shades of pink but on realising that I was using two colour schemes for the same meaning this changed. Encourage the eye to compare different pieces of data
All the data is laid-out in close proximity and values are colour coded. This allows for quick and easy comparison by glancing over the dataset.
A compare feature allows the user to study two pieces of data immediately next to each other. Mental strain is removed from the user by automatically calculating and then colour coding the difference between the two pieces of data (red for shrinkage, green for growth). This allows quick and accurate comparison. Reveal the data at several levels of detail
The comparison feature allows the user to reduce the detail they interact with to two pieces of data. Serve a reasonably clear purpose
There is introductory text in the main window which details the purpose of the visualisation. The snapshot window has legends and labels to provide context.
Accuracy of the representation of data and its information content All data points have been checked against the original data provided and that they have been plotted accurately on the visualisations. Test data was used to ensure the axes were not mal-aligned in the time series and that the snapshot box widths were uniform and accurate. Danny King (undergraduate of Computer Science at Durham University) [email protected] www.dannyking.eu
Page 1 of 2
13/04/2010
Danny King (kxrs26): Health Informatics Visualisation Introduction The software was created using Java and OpenGL (JOGL bindings). I enjoyed this assignment; it was interesting to apply the theory from lectures to a real problem and using OpenGL taught me a lot.
How to use the software To run the software run the DannyKingVisualisation.jar file, ensuring it is in the same folder as the gluegen-rt.dll, jogl.dll, jogl_awt.dll and jogl_cg.dll files. You will need a screen resolution large enough to display the largest window, which is 800 x 700px. You will be presented with the main screen, from which you can choose to view the time series data or the 2007 snapshot data. Rather than list the features you can interact with as bullet points I will label screenshots for clarity.
The time series window Interactive line graph showing the rising level of obesity in England.
Toggle emphasis of the maximum and minimum values for the 3 data sets.
Toggle data display on or off for total population, males or females
Watch the data evolve over the 11year period. One second represents a year’s data addition. You may interact with any of the other features whilst the animation is playing
Hover over any of the data points on a line to see its value appear next to it
The snapshot window Bar visualisation showing the relationship of long-term illness and waistcircumference
Select two of the data sets using the drop-down menus and click ‘compare’ for easier, side-by-side comparison & analysis
Danny King (undergraduate of Computer Science at Durham University) [email protected] www.dannyking.eu
For even easier comparison the difference between the compared data sets is calculated and displayed (green for a growth, red for a shrinkage)
Danny King (kxrs26): Health Informatics Visualisation 13/04/2010
Page 2 of 2
Thought that has gone into the time series data Edward Tufte’s principles of graphic excellence & graphical integrity
316 − 39 Graphical Integrity 39 𝐿𝑖𝑒 𝑓𝑎𝑐𝑡𝑜𝑟 = = 0.963 7988 − 954 Care has been taken not to distort the message of the data; the axes 954 are uniformly incremented and are not misleading in their starting Where 316 and 39 are the number of positions. Points on the graph are plotted at regular intervals. pixels above the x axis of the highest The lie factor is within Tufte’s acceptable range of 0.95 – 1.05 and lowest point of the ‘total’ line Induce the viewer to think about substance, not methodology, design, etc. respectively and 7988 and 954 are No unnecessary graphics, colours, effects or ‘ink’ were used and the the values of those two points from the dataset provided. visualisation has been kept as simple as possible. A line graph was used as most people are accustomed to them, enabling people to focus more easily on the data rather than the design. Encourage the eye to compare different pieces of data Toggling any of the total, male or female data sets on or off allows easy comparison between them. The ability to make the minimum and maximum points to stand out allows for quick comparison & analysis. The time evolution animation allows the user to see data changes over time, allowing easy comparison of how time affects the data. Reveal the data at several levels of detail The ability to toggle the 3 data sets on and off and also to view the data grow over time (evolution animation) allows for different levels of data to be interrogated at different times. Hovering over a point on the graph displays its value next to it, allowing for extra detail. Serve a reasonably clear purpose There is introductory text in the main window which details the purpose of the visualisation. The time series window has legends and labels to provide context.
Thought that has gone into the 2007 detail data Edward Tufte’s principles of graphic excellence & graphical integrity
95 − 61 61 𝐿𝑖𝑒 𝑓𝑎𝑐𝑡𝑜𝑟 = = 0.965 33.5 − 22.5 22.5
Graphical Integrity Where 95 and 61 are the pixel Again, care has been taken not to distort the message of the data; the widths of the female, raised waist widths of each rectangle are all equal. Any percentage value would circumference values for limiting have the same width in any rectangle. and non-limiting longstanding illness The lie factor is within Tufte’s acceptable range of 0.95 – 1.05 respectively (chosen for easier Induce the viewer to think about substance, not methodology, design, etc. calculation) and 33.5 and 22.5 are Again, No unnecessary graphics, colours or effects were used and it is the values of those two pieces of data from the dataset provided. as simple and compact as possible. The user’s eye is drawn immediately to the data which is predominant and in the middle. All colours have meaning. Originally the female data was in shades of pink but on realising that I was using two colour schemes for the same meaning this changed. Encourage the eye to compare different pieces of data
All the data is laid-out in close proximity and values are colour coded. This allows for quick and easy comparison by glancing over the dataset.
A compare feature allows the user to study two pieces of data immediately next to each other. Mental strain is removed from the user by automatically calculating and then colour coding the difference between the two pieces of data (red for shrinkage, green for growth). This allows quick and accurate comparison. Reveal the data at several levels of detail
The comparison feature allows the user to reduce the detail they interact with to two pieces of data. Serve a reasonably clear purpose
There is introductory text in the main window which details the purpose of the visualisation. The snapshot window has legends and labels to provide context.
Accuracy of the representation of data and its information content All data points have been checked against the original data provided and that they have been plotted accurately on the visualisations. Test data was used to ensure the axes were not mal-aligned in the time series and that the snapshot box widths were uniform and accurate. Danny King (undergraduate of Computer Science at Durham University) [email protected] www.dannyking.eu
