Many ways to skin a cat but please don't ring 999...
In a few past posts I explored the data on London Fire Brigade incidents, and given the wealth of interesting informtation it contains (more for identifying anecdotes for the pub than anything else) I may return to the theme again.
However, what really got me thinking was the number of ways that you can present data to try and get a message across. Very often a table is the easiest, albeit least sexy way of doing this. However, when there's a geographic dimension, I think a map can be very power if done well, but one's ability to do this may be constricted by available software. However, with a little time and patience, and using what you do have available, you can start developing ideas.
Initially, I used conditional formatting on data from a pivot table, which gave me a rather nice 'mosaic' map of Tower Hamlets, but is limited in that the colours reflect the the underlying value within a band.
I then presented my data table in a slightly different way to be able to produce this bubble chart, which shows the relative number of incidents dealt with in a given area showing scale. However, because the bubbles overlap, things get a bit cluttered, but this could arguable aid with understanding 'hot spots' (or perhaps cold spots to a greater extent) and the division of the data into 4 series for each quartile, allows some more insight.
I then tried surface charts, which allow each 100m square to be displayed without any overlap, with number of incidents shown vertically. This presentation is pretty useless for conveying how many incidents occured in a given place, but does show nicely the massive range in frequency across the area, though the data is perhaps a little too complicated, and Excel's ability to customise the chart makes this a very fiddly option.
After exchanging a few emails with LoveWapping, I ultimately pulled my finger out and converted the geographic element of the data into google compatible coordinates and I was able to produce this overlay on a google map which also shows where the fire stations are. This presentation works nicely because it is dynamic and allows the user to zoom in, but most importantly combines the data with the underlying geographical features and allows greater insight into the data, but clearly for presentation in a document would look quite unprofessional to have the very identifiable google maps beneath it (not to mention issues of copyright).
Hopefully though, what this does show is that there are many ways of visualising data that even I, a humble accountant can try to use to get information across.
However, what really got me thinking was the number of ways that you can present data to try and get a message across. Very often a table is the easiest, albeit least sexy way of doing this. However, when there's a geographic dimension, I think a map can be very power if done well, but one's ability to do this may be constricted by available software. However, with a little time and patience, and using what you do have available, you can start developing ideas.
Initially, I used conditional formatting on data from a pivot table, which gave me a rather nice 'mosaic' map of Tower Hamlets, but is limited in that the colours reflect the the underlying value within a band.
I then presented my data table in a slightly different way to be able to produce this bubble chart, which shows the relative number of incidents dealt with in a given area showing scale. However, because the bubbles overlap, things get a bit cluttered, but this could arguable aid with understanding 'hot spots' (or perhaps cold spots to a greater extent) and the division of the data into 4 series for each quartile, allows some more insight.
I then tried surface charts, which allow each 100m square to be displayed without any overlap, with number of incidents shown vertically. This presentation is pretty useless for conveying how many incidents occured in a given place, but does show nicely the massive range in frequency across the area, though the data is perhaps a little too complicated, and Excel's ability to customise the chart makes this a very fiddly option.
After exchanging a few emails with LoveWapping, I ultimately pulled my finger out and converted the geographic element of the data into google compatible coordinates and I was able to produce this overlay on a google map which also shows where the fire stations are. This presentation works nicely because it is dynamic and allows the user to zoom in, but most importantly combines the data with the underlying geographical features and allows greater insight into the data, but clearly for presentation in a document would look quite unprofessional to have the very identifiable google maps beneath it (not to mention issues of copyright).
Hopefully though, what this does show is that there are many ways of visualising data that even I, a humble accountant can try to use to get information across.
Comments
Post a Comment
Please leave a comment to let me know if you think I've got it totally or partially wrong!