(wip) visualizing federal marginal tax rates

cleaningFedData.py

'''cleaning data for Feds
   to execute, in correct directory type "python3 cleaningFedData.py"

     - for when I'm ready to do other categories
     myarray.append([year, row[5], row[6], row[4], 'married filed seperately', 'nominal'])
                myarray.append([year, row[9], row[10], row[8], 'single', 'nominal'])
                myarray.append([year, row[13], row[14], row[12], 'head of household', 'nominal'])
'''
import csv

def readList(filename):
    ''' reads in a file line by line
        for each line it strips the new line character, turns it into an integer, and appends it to an array
    '''
    line_number = 0
    myarray = [['tax_year', 'income_floor', 'income_ceiling', 'marginal_tax_rate', 'category', 'value']]
    year = 0
    with open(filename, encoding='utf-8') as a_file:
        for a_line in a_file:                                             
            line_number += 1
            row = a_line.rstrip().split(',')
            if (row[2] == 'Nominal'):
                year = row[6]
            elif (row[0].split('.')[0] == '0'):
                myarray.append([year, row[1], row[2], row[0], 'married filed jointly', 'nominal'])
    return myarray


if __name__ == '__main__':
    mylist = readList('FedDataToClean.csv')
    with open('parseddata.csv', 'w', newline='') as fp:
        a = csv.writer(fp, delimiter=',')
        a.writerows(mylist)

fed_income_tax.csv

fedTaxes.css

.focus text {
  font-family: Times New Roman;
  font-size: 24;
  fill: orange;
}

div.tooltip {   
  position: absolute;           
  text-align: center;           
  width: 170px;                  
  height: 14px;                 
  padding: 2px;             
  font: 12px sans-serif;        
  background: white;   
  border: 0px;      
  border-radius: 8px;           
  pointer-events: none;         
}

fedTaxes.js

/* 

  for mouseover, learning from this bl.ock: http://bl.ocks.org/mbostock/3902569
next steps: 
   - manipulate data efficiently to include all years
   - show detailed information on hover
   - legends/axis
   - on click, large pretty text saying 
          "if you made between X and Y in year Z, you would have paid k% in federal taxes (X' to Y' dollars)"
 
 */

// set variables for use throughout
var barHeight = 5;
var income_factor = 2000;
var boxsize = 1000
var colorLow = '#ffffe5';
var colorHigh = '#004529';

// set up svg
var svgHolder = d3.select("body")
	.append("svg")
	.attr("width", boxsize)
	.attr("height", boxsize);

// set up div for tooltip
var div = d3.select("body").append("div")
	.attr("class", "tooltip")
	.style("opacity", 1e-6);

/*
var focus = svgHolder.append("g")
	.attr("class", "focus");
//	.style("display", "none");

focus.append("text")
	.attr("x", 9)
	.attr("y", 0)
	.attr("dy", ".35em");
	*/

// build color scale from low to high marginal tax rate
function setColorScale(max) {
	return d3.scale.linear()
		.domain([0, max])
		.interpolate(d3.interpolateRgb)
		.range([colorLow, colorHigh])
}

// 
function drawRectangles(svg, mydata) {
	var maxTax = 0;
	var maxIncome = 0;

	// find max tax rate and max income in dataset
	// change 0 to 1 for income floor, so can use log scale
	mydata.forEach(function(d) {
		if (+d.marginal_tax_rate > maxTax) {
			maxTax = +d.marginal_tax_rate;
		};
		if (+d.income_floor > maxIncome) {
			maxIncome = +d.income_floor;
		};
		if (+d.income_floor == 0) {
			d.income_floor = 1;
		}
	})

	// replace 'na' with max of dataset; maybe will preprocess instead?
	mydata.forEach(function(d) {
		if (d.income_ceiling == 'na') {
			d.income_ceiling = maxIncome;
		};
	})

	// set scales
	colorScale = setColorScale(maxTax);
	incomeScale = d3.scale.pow().exponent(.2).domain([1, maxIncome]).range([0, boxsize]);
	console.log(maxIncome)
	yearScale = d3.scale.linear().domain([1862, 2013]).range([500, 10]);

	// draw rectangles
	svg.selectAll("rect")
		.data(mydata)
		.enter()
		.append("rect")
		.attr("x", function(d, i) {
			// x start is determined by income floor
			return incomeScale(+d.income_floor);
		})
		.attr("y", function(d, i) {
			return yearScale(+d.tax_year)
				//return (2013 - +d.tax_year) * barHeight;
		})
		.attr("width", function(d) {
			// width is difference between scaled floor and scaled income ceiling values
			return incomeScale(+d.income_ceiling) - incomeScale(+d.income_floor)
		})
		.attr("height", barHeight)
		.attr("stroke", 'white')
		.attr("fill", function(d) {
			return colorScale(+d.marginal_tax_rate);
		})
		.on("mouseover", function() {
			div.transition()
				.duration(500)
				.style("opacity", 1);
		})
		.on("mouseout", function() {
			div.transition()
				.duration(500)
				.style("opacity", 1e-6);
		})
		.on("mousemove", function(d) {
			console.log(this);
			var coord = d3.mouse(this);
			var income = incomeScale.invert(coord[0]);
			div
				.text('year: ' 
					+ Math.ceil(yearScale.invert(coord[1])) 
					+ ', income: ' 
					+ d3.formatPrefix(income).scale(income).toFixed() 
					+ d3.formatPrefix(income).symbol)
				.style("left", (coord[0] + 20)  + "px")
				.style("top", coord[1] + "px");
			/*	focus.attr("transform", "translate(" + coord[0] + "," + coord[1] + ")");
		focus.select("text").text('check this out'); */
		})
var prefix = d3.formatPrefix(13759402);
console.log(prefix.symbol); // "M"
console.log(prefix.scale(13759402).toFixed());
}

d3.csv("fed_income_tax.csv", function(error, dataset) {
	drawRectangles(svgHolder, dataset);
});

index.html

<!doctype html>

<html lang="en">
<head>
  <meta charset="utf-8">

  <title>Federal Income Tax Rates</title>

  <link rel="stylesheet" href="fedTaxes.css">

  <!--[if lt IE 9]>
  <script src="http://html5shiv.googlecode.com/svn/trunk/html5.js"></script>
  <![endif]-->
</head>

<body>
  <script src="http://d3js.org/d3.v3.min.js" charset="utf-8"></script>
  <script src="fedTaxes.js"></script>
</body>
</html>