Writing by Numbers: an index of statistical prose

Have you heard of the Harper's Index? Often described as statistical poetry, the monthly index highlights the numbers behind the stories. A bare scaffolding is exposed and we are the richer for engaging.

The July 2018 index below begins with a thematic discussion of self-storage units and by creative association weaves vertically down the page to culminate in NYC sewage traveling by train. Incrementally this makes sense if you follow the thread.

---

Estimated combined square footage of all the self-storage units in the United States : 1,670,000,000
Factor by which this area is larger than the borough of 
Manhattan : 2.6
Number of states that prevent cities from enacting rent control laws : 37
Amount a US company is charging for an eleven-night stay at a luxury space hotel scheduled to open in 2022 : $9,500,000
Number of reservations made during the first month bookings were available : 22
Percentage of first-time US home buyers who are single men : 7
Who are single women : 18
Factor by which a mother who has been laid off is more likely to get a job interview than a stay-at-home mother : 2
Chance that an ad for a Chinese central-government job specifies a preference 
for men : 1 in 5
That it specifies a preference for women : 1 in 31,733
Percentage of Americans who consider themselves members of the “alt-right” : 3
Who have never heard of the alt-right : 5
Percentage of the Brazilian population that is black or mixed race : 55
Of sperm imported into Brazil that is from white donors : 95
That is from blue-eyed donors : 52
Months for which Denmark will require couples with children to “reflect” before finalizing divorce : 3
Number of countries in which divorce is outlawed : 2
Percentage by which more EU nationals have arrived in than left the United Kingdom since the Brexit vote : 70
Date on which the United Kingdom announced a return to 
blue-colored passports : 12/22/17
Value of the contract for the passports, which was awarded to a French-Dutch company : $353,000,000
Amount of public money Mexico has spent since 2013 on ads promoting government accomplishments : $2,000,000,000

Portion of money spent by Donald Trump’s reelection committee this year that has gone toward legal fees : 1/5
Percentage of Trump voters who now say they regret their vote : 3
Of Hillary Clinton voters : 3
Factor by which an American between 18 and 29 is more likely to be fearful than hopeful about the future of the country : 2
Number of successive years that health care has been the political issue that Americans care most about : 5
Portion of high-net-worth US investors who regard health care as their greatest financial concern : 7/10
Factor by which the largest pension in Oregon is greater than the average Oregon household income : 11
Rank of 2017 among years in which US bankers received their highest average bonuses : 2
Of 2006 : 1
Average amount an Australian loses gambling each year : $957
Percentage of Australian state and territory tax revenue that comes from gambling : 7.7
Total amount owed by Pennsylvania’s top 100 evaders of highway tolls : $3,400,000
Number of the ten largest US public transit systems that saw an increase 
in use last year : 1
Projected percentage increase in energy consumption from 
cooling appliances by 2050 : 90
Percentage chance that a bottle of water contains microplastic particles : 93
Estimated portion of European fish species that a planned series of dams in the Balkans would put at risk of extinction : 1/10
Number of US states that ban the sale of fake urine : 21
Percentage of New York City’s sewage that is shipped out of the city for disposal : 85
Number of weeks a train filled with New York City sewage was stuck outside an Alabama town this year : 11

---

This wonderfully orchestrated index got me thinking. What if this format--loosely reimagined--could serve as a guide to data exploration? I guide many companies and professionals along the path of data literacy. Typically, I create a few visualizations and we spend the afternoon unpacking what's under the hood. I am not a big fan of a traditionally didactic approach. I view data as dynamic and evolving. I know, I know--the headlines herald the arrival of artificial intelligence and machine learning as the panacea to insights and innovation.

But most of us are small data folks. Somehow the data explosion left many voiceless in a world where the esoteric terminology arrived without a Rosetta Stone. Even if we have a serviceable amount of data literacy, where do we find the data to address our questions?

I rely on the American Community Survey and Census data when researching social determinants of health. I have a strong interest in race data because of my own curiosity regarding identity but also professionally. How can strong statements about race be made in the absence of wide datasets that capture actual genetic variants or social correlates?

So I started thinking of my own story. My plan was to find a data stream not unlike the Harper's Index and share the data sourcing to hopefully encourage others to create a data catalogue of useful information. I decided to begin with the census report applicable to my birth year. The Census data is where you can find population data relevant to discussions of social correlates of health in addition to your own lingering questions about the world we live in.

Think of the data journey possible anchored by statistics centered on Negro Population reported in the 1970 Census. As a bi-racial woman even I was startled to see the historic language prevalent in what seems like a not-too-long ago government document. Where would my index begin?

---

Percent distribution of negroes reported residing in the Northeast in 1969--19%

The year in which interracial marriages were deemed legal by the U.S. Supreme Court--1967

​Percentage of all marriages which were racially mixed in US (1967-1970) [specifically NJ where I was born]--0.6%

*data from a 1975 funded project by National Institutes of Health #1-RO1-HD-05137

Census by Decades

Additional sources of population totals by race 1790-1990 available here --> Historical Census Statistics on Population Totals by Race, 1790 to 1990, and by Hispanic Origin, 1970 to 1990, For Large Cities and Other Urban Places in the United States

---

Are you interested in data sourcing to understand population demographics and communities and how publicly available data sources can be accessed to fill in the gaps?

Follow along here or twitter @datamongerbonny