Having worked with visualization of data for some time, I’m always looking at new ways to express data into a meaningful representation, either by restating it through visualizations, aggregations, abstraction, or introducing new data sets to complement it.
Whilst I can’t discuss much of my work due to the confidential nature of the data, there is public data out there which is readily available and very insightful. In the visualizations below I used data from the US Dept. of Labor (Labour) whose census recorded the hours worked across different industries
What stands out most clearly is Healthcare, Protect Services (Security, etc), Food services and Transportation industry workers cover more of the day (and night) than most other industries. Of course this might not come as a shock to many of you (especially if you’ve worked in those industries), but interestingly it also shows that “blip” of the lunch hour of noon – and farm/fishery workers are some of the most prolific “non-workers” during that time.
For added measure I also included another “map” visualisation – whilst not as detailed as the previous heat map, it demonstrates how we can look at data in a number of ways. In effect, depending on the question we’re looking to answer we might want to aggregate information and pick out top/bottom 20% (for example). In the case of this visualisation, we’ve come to the same conclusion but possibly avoided the need for people to see lots of % numbers per “cell” (as in the first visualisation), which can be distracting. This particular visualisation could be a great “opener” and then invite users to click through to the most interesting data points, rather than cause more decision from the user if they were to see the first (e.g. they start to evaluate what their threshold is – 20%? 25%?, etc.).
For added visualisation “fun”, here’s some more public data correlation – Illinois unemployment with retail gas prices and crude oil prices – http://public.tableausoftware.com/views/GasandUnemployment/CorrelationofGasUnemployment?:embed=y