First, here are the most interesting trends that I extrapolated from

Health and Family

  • > 50% of the population will be obese in 35 years (currently 29.8% of the population)
  • Cancer deaths has grown at the same rate as population growth (1% YoY)
  • Life expectancy will be 118 years of age in 100 years (78.8 years right now)
  • Divorce rate will jump from 15% a year to 25% in 50 years


  • Armed forces will be less than half the size it was in 1980 in 28 years (if there’s a place for robots and machine learned automation, it’s in the military then – and more data-driven targeted recruiting campaigns)


  • Only Federal / State / Local ran business that loses more money each year (not adjusted) is State / Local Transit. Without Transit, State / Local Business would be positive by $15B in 2015, and all Government-run businesses would be positive by $12B vs. $39B net negative. In fact, if State / Local Transit continues to grow at the same rate for the next 12 years, it will be $100B net negative (not adjusted). When we compare the growth rate of population and transit net expenditures since 2010, transit loss outpaced population growth by 4X.
  • In 2022 (eight years), US Government revenue will exceed expenditures. This is modeled based on the growth for expenditures and revenues over the previous 10-year period from 2004 – 2014 – covering five years of George W. Bush’s term and five years of Barack Obama’s term. Which begs the question – why aren’t US presidents evaluated based on future growth (are they improving the forecast or not)? The president who inherits the period of surplus will likely reap the full credit, even though gradual changes needed to happen over many years to result in that surplus.
  • In 1980, the economy generated one net new job for every 106M people in the population. Currently, it’s one net new job for every 117M people. In 50 years, it will be one net new job for every 146M people (a difference of 40M people not being addressed per net new job compared to in 1980). Net new jobs is declining 0.4% YoY (while the population is growing 1% YoY).


  • In two years, the US will generate more barrels of oil in a year than Saudi Arabia (10,608 vs. 10,051 thousands of barrels per day).
  • In five years, US will be able to produce more energy than what the US consumes presently. In four years, US will produce more energy than what is expected to be consumed in the US in that year (energy consumption is declining 0.28% YoY).


Last week, Steve Ballmer launched, a non-profit, one-stop government data service. Steve has invested $10M into the project, and after exploring the site in great depth this past week, I believe he made a wise investment given the considerable value that this service provides.

The site is well designed and hosts many beautiful visualizations of government data – data that has been effectively inaccessible as they’re buried in hard to find government webpages, and presented or saved in old school HTML tables or Excel workbooks lacking any standards around table formatting, date / timestamps, etc. The USAFacts team must have spent months just wrangling all of this data from the numerous, disparate government data sources.

I spend most of my time these days on enterprise AI and analytics software, and the design, filters, search capabilities, and of course, data visualizations of USAFacts blow away any dashboards that I have seen inside companies. I especially like how USAFacts grouped the data and visualizations by mission – missions gleaned straight from the US Constitution.

The USAFacts project deserves praise, but it is certainly an early beta product. The site seems more focused on aesthetics for “window shoppers” versus interactivity to enable users to dig into the numbers, as the site currently does not expose the sources or the raw data behind many plots.

Additionally, the charts are fairly basic and do not correlate trends with other sources (for ex. S&P 500 prices), and several statistics really ought to be normalized by population size or adjusted for inflation.

I also noticed rounding issues. For ex. USAFacts claims that the US produced 3,437,000 barrels of oil in the most recent recorded year, but if you cross check that with and take their 9,415 thousands of barrels per day figure * 365 days in a year = 3,436,475 barrels. The numbers are close, but exactness is necessary especially for forecasting trends as small rounding issues lead to bigger deviations over time with years of compounding.

This is why it would be nice to see the sources and the underlying raw data to verify their assumptions. This should be expected from fact checking services, and I would have prioritized this for the first version release at the sacrifice of visualization prettiness.

There is still so much to like here though – and it is just a first version that is only weeks old. Steve also seemed to go out of his way to try to make this work as impartial as possible – he didn’t even take a tax deduction for this non-profit work in order to avoid stirring up any possible bias (not that he would need the deduction).

Steve put his name on this project and sold it well via PR. They wrangled together data that is difficult to access. They presented and hosted the data via a beautiful web service. This is how you market and bring data to life. Imagine if other celebrities invested in data sets (climate, health care, etc.), and brought them to life like Steve Ballmer did with USAFacts. When this happens, it encourages others to dig into the data to become informed – like I did with USAFacts.

I understand the motivation to remove political bias by just revealing the raw statistics versus drawing conclusions, as USAFacts sets out to do. However, we still need people to transform these raw data points into insights to drive action at the risk of being imperfect or biased, or else, what’s the point of collecting the data? The trends that I uncovered in just a few hours from USAFacts are eye-opening, and I hope to see more efforts to analyze and open up more data sets that can help us better understand, and thus improve, the quality of life for everyone.

Sources and Calculations


((29.8/20.1)^(1/15))^35 * 20.1


pg. 39



(321/226)^(1/(2015 – 1980))

Life Expectancy

78.8 * ((78.8/73.7)^(1/(2013-1980)))^118


15 * ((15/12)^(1/(2015 – 1993)))^50

Armed Forces

2061662 * ((2061662/3010743)^(1/(2013-1980)))^28


((51.13/13.18)^(1/(2014-1990)))^12 * 51.13

Government Revenue and Spend


5217217757000 * ((5217217757000/3323158594000)^(1/(2014-2004)))^8


5385587977000 * ((5385587977000/3582440073000)^(1/(2014-2004)))^8

Net Change in Jobs

Net change in jobs going down 0.4% a year

(2-2.471/2.741)^(1/(2016 – 1994))

Population growing almost 1% a year


322/2.741 = 117M people in population per net new job now

263/2.471 = 106M people in population per net new job in 1980

146M people to each net new job in 50 years

117 * ((117/106)^(1/(2016 – 1994)))^50

Oil Production


3437000/365 * ((3437000/1892000)^(1/(2015-2005)))^2

10,610 thousands of barrels per day

Saudi Arabia

10046 * ((10046 * 365)/(9550 * 365))^(1/(2015-2005))^2

10, 051 thousands of barrels per day

Energy Production and Consumption

In five years, US will have an energy surplus compared to most recent energy consumption figure

88020 * ((88020/69440)^(1/(2015 – 2005)))^5

In fact, surplus will happen in four years when compared to projected energy consumption in four years


88020 * ((88020/69440)^(1/(2015 – 2005)))^4


97330 * ((97330/100190)^(1/(2015 – 2005)))^4