Hi, let's talk about numbers and data and how to make images from things like that so people can understand what the heck. 

KNOW THY DATA:

the basic building block

all things flow from here.

 

 

 

 

 

 

 

numbers simply describe the world

let them.

 

 

 

 

 

 

 

making charts is not so hard, finding what to make charts of can be a great challenge.

 

 

 

 

 

 

 

 

 

 

what i do-

spreadsheets are how machines read- but terrible for us.

we (PEOPLE) are pattern recognition machines- show the patterns

 

 

 

 

 

 

 

 

a small piece of advice:

GARBARGE IN, GARBAGE OUT

 

 

 





a small piece of advice:

GARBARGE IN, GARBAGE OUT




 

 

 

 

formats- xls, csv, tsv, sql, shp, 

the dreaded PDF

 

 

 

where to find-

government websites

specific websties for your topic

FIOA requests

ASK YOUR SOURCES

roll your own - 

http://tmsnrt.rs/2gKl5Ro

https://podio.com/site/creative-routines

vs

http://static1.squarespace.com/static/56ad0eb962cd942d9e46f366/56b9123cf04e930f40f0c17c/56b921c8f04e930f40f17ddb/1454973384243/creative-routines-edit3.png?format=original

 

 

 

 

 

data is plural - good resource

explore! start a folder, bookmarks, notes, etc

follow the trails- look at bottom for sources, everyone has data, assume they do until they tell you they don't

 

 

 

 

analysis-

for larger datasets, look for databases / ask for data dictionary- KNOW THY DATA

 

statistics = sexy.

don't kid yourself.

 

 

 

 

a few thoughts on interpreatation:

 

correlation is correlation.

causation is causation.

causation is usually (linear) correlation.

but C'MON NOW

correlation is not causation!

don't get crazy

 

http://www.tylervigen.com/spurious-correlations

 

 

 

 

tools:

excel,

googlesheets,

sql,

sublime text,

GIS

​online what have you

 

google is everything.

 

 

 

 

watch for:

naming conventions, date formatting, number formatting, column names

states, business names, county names, FIPS codes

 

 

 

cleaning issues

make things uniform to save yourself pain.

a lot of this comes with massive amounts of repitition, knowing what to look for, the warning signs.

 

 

once you have your data ready, now comes the hard part-

what does it say? how to best present?

familiarize yourself with the range of charts-

 

http://datavizproject.com

 

 

 

 
presentation:
DO NOT BE TAKEN IN BY RAZZLE DAZZLE
good data makes an impact simply by it’s shape and the shape of it’s ideas.
 
if you have to explain why it’s good, it’s not good. it should present on it’s own.
people do not read charts!
 
 
 
 
 
never:
dual axis
bar charts always start at 0.
bar vs line- a common conundrum
it’s not always a map.  think it's a map? it's probably not
pie charts- meh
hard to compare slices, overall size often done incorrectly.
(area not just radius = pi*rsquared)



never:
dual axis
bar charts always start at 0.
bar vs line- a common conundrum
it’s not always a map.  think it's a map? it's probably not
pie charts- meh
hard to compare slices, overall size often done incorrectly.
(area not just radius = pi*rsquared)
 
 
your three part mantra:
data
analysis
story
 
all presentation decisions flow from these
 
ask yourself questions
when confronted with
head scratching
or hard decisions.
 
 
 
Then, as you would choose words and adjectives to frame ideas, you use colors and bars and lines and circles to clearly illustrate concepts with rigor and panache.
 
easy.
 
 
 
 
 
 
people do not read charts!






people do not read charts!
 
 
 
 
 
 
COLORS!
highlight one.
 
is it a MAP?
probably not.
 
BAR VS LINES
rate?
total amount of a thing? 
percent change?






COLORS!
highlight one.

is it a MAP?
probably not.

BAR VS LINES
rate?
total amount of a thing? 
percent change?
 
 
they are an entry point.  eye magnets
 
 
 
make it count.
there are many details to consider.
 
 
type sizes-
fonts-
be simple.
simple
simple
reduce to the bare essence.
be poets.
 
Don't go on forever blathering about what ever nonsense you think is important, if it's not.  ask the data what it needs to be and let it be that.  if the data and your story aren't seeing eye to eye, let it go away. Then go find better data.
how to make a chart for medium in like 7-109 easy steps.
 
 
 
 
 
 
 
 
 
 
backup link in case everything goes to hell
https://public.chartblocks.com/c/5a1382ea9973d2e511e88d61?t=008660f668e1c46
stacked bar:http://tmsnrt.rs/2nISh9x
stacked bar
stacked dots :http://tmsnrt.rs/1WQw1IP
stacked dots
small multiples: http://tmsnrt.rs/24xuzOM
small multiples
scatter plot: http://tmsnrt.rs/2boaHHZ
scatter plot
lights out - http://tmsnrt.rs/2us7qDM - mapping in three
lights out - mapping in three
venezuela - http://tmsnrt.rs/2pPJdRb - 
venezuela;
protests - http://tmsnrt.rs/2ujuylf
protests
opoiod: http://tmsnrt.rs/2hO4YC7
opoiod
Made with Slides.com