So here is a Tableau visualization of all the Obamacare health plans for states that elected to use the federal exchange (not sure about the states where the Feds and state Gov’t split some responsibility). Here is a breakout map. The values differ by county so I am showing the averages. Full visualization of all the health plans can be found here.
Clicking on any circle, or highlighting a section will bring up those plans below. You can then sort and filter as needed.
Interesting dashboard I made here showing Average (not median) pay for each company, and the pay for their CEO. If you will notice, the CEO of J.C. Penny’s was paid an exorbitant amount! What the heck J.C. Penny, why did you pay Ex CEO Ron Johnson so much? HE WAS 6 STANDARD DEVIATIONS FROM THE NORM! 6!!! WHAT THE HECK!
Here is a tenure change report that I did awhile ago in a data viz class. I wanted to post it because I think the page shelf in Tableau is under-utilized. All the pages are shown at once by using the ‘show history’ option in Tableau. This example here shows how tenure in some positions have changed, but others have not. I have two graphs showing the same thing, but in the chart at the bottom I have turned everything to the right. Here are some tips I’ve stumbled upon:
A dotted line seems to show ‘something is changing’ better than a solid. It shows impermanence.
The past tends to fade. Fade (adding transparency or whiteness) seems to indicate the past better than a change in hue.
Keep the change 1 dimensional. After all our dimension of time is one dimensional (and so far one directional). Sometimes you can get away with 2 though as Hans Rosling has done in http://www.gapminder.org/
Up does not mean the same as Down, Left, or Right.
In this chart, I have deliberately chosen Up rather than right.
2nd chart: Notice how everything looks ‘faster’
Although the 1st graph has a wider professor category area, the Number of Positions axis on both are relatively close in size. (Please note that you need to measure to 160k, I forced it to 180k on the 2nd graph to get the label in). However the small changes in perspective change this graph completely.
Recently I read an article from the AJC showing the top paid GA State employees. Their article can be found here. (note that at the time of this publishing their website does not link to the article). I have made a dashboard showing who gets paid the most by university. Travel and Salary expenses are included. First of all,this excludes the coaches, who get paid a large amount. Secondly if the person started the position half-way through the year then their pay is only 1/2 what it should have been (i.e. Savannah State’s President). Third thing to know is that there was no Part-Time or Full-Time designation in the data I received. As such you will have to decide on an arbitrary ‘part-time cut off’ using the slider at the top. I have decided on a $10,000 level. You can put in $15k if you wish. It’s really not certain what part-time is. This could make a huge difference.
The ‘ratio’ column is the ratio between the highest-paid person and the average pay for the university (again that average pay is dependent on the slider).
The graph shows the university’s average pay, median pay (both based on the filter), max pay, and the president’s pay. The person’s name to the right is the highest paid person for that university, If you hover over the colored box next to their name you will see their position.
All the data came from OpenGeorgia.gov. It includes Travel and Salary. One special thing to note is that apparently some people may be paid by the university foundations and that pay may not be included in this dashboard.
This article in Wired seems to imply that there is a correlation between defense/intelligence industry contributions and whether the Congressman voted to continue NSA funding. Lets be clear, the amendment had a snowball’s chance in a Georgia July (although what the heck GA this summer has been very mild!) of passing through the Senate and Obama’s veto, but Wired had some data section detailing the 2 year contributions and the result of the vote.
I first took the data, sorted it by the contributions from largest to smallest, and gave a -1 for a ‘No’ vote a 0 for a ‘not voting’ and a 1 for a ‘Yes’ vote. A “no” vote is a vote to continue the NSA’s phone spying. I then did a running total and produced this graphic in Excel:
Just eyeballing it, it seems to me that down to ~$23k in contributions (I think over 2 years) a congressman was more likely to vote to continue funding domestic surveillance and around $18k and lower a congressman was much less likely to vote for it, but not a guaranteed no. There were still some yes votes in the low-contribution range. Note that the axis is a discrete listing of all contributions and note a range.
I produced another quick chart to breaking each congressman into deciles based on campaign contributions.
It would seem to me that although contributions appear to be a deciding factor for voting yes in some of the higher deciles, getting fewer dollars in contributions was by no means a contributing factor in voting yes or no for the bulk of the congressmen. (Hey maybe those guys getting 0 dollars really want it, so maybe that’s why they voted yes). Also note that these $ amounts don’t take into effect the total contributions given to a candidate. It is likely that percentage matters and I didn’t have time to collect all the data.
First of all, sorry for the delay in posting. Life has been very different in these last 6 months. I have a backlog of ideas I need to work through. Anyways a few weeks ago I read this article from Zero Hedge titled “Spot The Grotesque Retail Sales Seasonal Adjustment Outlier” and it seemed to imply that the Government has grossly over/under adjusted seasonal retail numbers for June. While I admit that the original article does seem to make a good point with the outlier, I wanted to provide more context than their original chart.
For reference here is their original chart:
I have provided some other charts using Tableau to show how this June adjustment is not very remarkable (although the sign change from – to + is interesting).
In this first graph I am showing all the adjustments by month for all the past few years. As you can see June (in red) is hardly as dramatic of an outlier compared to other months:
In this second chart I have plotted the log of the difference in adjustments and I have labeled the min and maximum years. (Without the log of the difference 1992 seemed to show up more).
So here is the entire view. You can see that other months ‘straddle’ the zero seasonal adjustment axis, and June ‘was always close’ to the zero line, so it is ‘conceivable’ that there is nothing wrong with June 2013′s adjustment. However even at this larger view, June 2013 does seem to stands out! Is it a large deviation? No not really considering the whole picture. This is either a legitimate change in American retail sales for June (is Memorial day becoming a larger retail event than in the past?) or a one-time adjustment (June 2013 weather event?). I have no opinion, but I wanted to present a different picture.
I have been playing with some sankey graphs. This graphic shows the change in labor costs by departments year over year. This is not per person, just the total labor costs by department. Please note that it excludes certain educational employees, read the fine print in the graphic.
Using Slice by Juice Analytics I was able to make these two dashboards showing state employee pay in Georgia. I collected the data, scrubbed some of the names and departments and excluded anyone with less than $1000 avg pay in the last 4 years. Data was collected from http://www.open.ga.gov and includes only State Agencies,Boards,Authorities and Commissions. Units of the University System and Georgia Military College, Regional Educational Service Agencies, Technical Colleges, and Local Boards of Education are excluded.
Keep a few things in mind as we look at this data. The Teachers retirement system has ~55 billion in assets. The GA Lottery has around 38 billion in sales. The GA ports authority does 26.5 M tons annually.
I have made another program for the blink(1) product. This Processing sketch uses the mouse location and determines the color of the blink(1) light. The speed of the mouse will determine whether the light blinks. The ‘blink threshold’ will determine what distance in pixels the mouse must travel before it blinks. Press ‘b’ to lower this and ‘B’ to increase it. If your mouse moves more than X pixels per second then it will blink. The interface for the program currently looks like this:
There is 1 bug I have discovered. When windows 7 logs you off (or you press ctrl+alt+del) the program will crash.
So I finally got my blink(1) from ThingM! I will call the blink(1) product a small-scale indicator. Small-scale sensors have been proliferating in recent years. The smartphone is the most obvious. Another revolution I see coming is in the small-scale indicators. Blink(1) is one such device. Since I love Processing and anything nerdy, I have written a Processing Sketch to show some of the abilities of blink(1). The Mouse’s X and Y position determine color. The blinking is based on the speed of the movement. After a certain pixel distance threshold the color is set to 0.
First you obviously need to get a blink(1) and plug it in!