DSP Engineer turned Data Scientist @ccri
Published
Edited
May 22, 2021
2 forks
89 stars
arrow = require('apache-arrow@4.0.0')
scrabble = arrow.Table.from(fetch(url))
scrabble.schema.fields.map((d) => d.name)
scrabble.get(0).losername
function columnStats(name) {
// Get the column withe the given name
const column = scrabble.getColumn(name);
// Use it's iterator to find the max and min values
let max = column.get(0), min = max;
for (let value of column) {
if (value > max) max = value;
else if (value < min) min = value;
}
return {max, min, range: max-min};
}
winner_table = scrabble.countBy('winnername')
winner_table.toJSON()
biggest_winners = collectCounts(winner_table).slice(0, 30)
losers = {
const counts = scrabble
.filter(arrow.predicate.col('winnername').eq(winner))
.countBy('losername');
return collectCounts(counts);
}
// Find the winner in the winnername column's dictionary
winner_key = scrabble.getColumn('winnername').dictionary.indexOf(winner)
{
// Look for the winner's key in the indicies and gather all the matches
let result = [], idx = 0;
for (let key of scrabble.getColumn('winnername').indices) {
if (key == winner_key) result.push(idx)
idx++;
}
return result
}
// Render scores from a DataFrame (df) onto an ImageData (image) with a given color (32-bit ABGR)
function plotScores(df, image, color) {
const h = image.height, w = image.width,
image_buf = new Uint32Array(image.data.buffer);
let wscore, lscore, x, y;
df.scan(
// `next` function
// Uses the wscore,lscore accessors assigned in `bind`
(idx) => {
// Map the winner score to a y value
y = h - Math.round((wscore(idx) - stats.winnerscore.min)/stats.winnerscore.range*h);
// Map the loser score to an x value
x = Math.round((lscore(idx) - stats.loserscore.min)/stats.loserscore.range*w);
// Fill that pixel in the image buffer
image_buf[y*w + x] = color;
},
// `bind` function
// uses col's .bind() method to get accessors for winnerscore and loserscore
(batch) => {
wscore = col('winnerscore').bind(batch);
lscore = col('loserscore').bind(batch);
}
);
}
all_scores = {
// Max out at 1 pixel per point
const w = stats.loserscore.range < width ? stats.loserscore.range : width;
const pixels_per_point = w/stats.loserscore.range;
const h = stats.winnerscore.range*pixels_per_point;
const canvas = DOM.canvas(w, h);
const ctx = canvas.getContext("2d");
const image = ctx.createImageData(w, h);
plotScores(scrabble, image, 0xFF000000);
ctx.putImageData(image, 0, 0)
return canvas;
}
filtered = scrabble.filter(col('year').eq(year_filt))
{
const h = all_scores.height, w = all_scores.width;
// Grab the image buffer from "all_scores"
const image = all_scores.getContext("2d").getImageData(0,0,w,h);
// Add some blue dots for the currently selected year
plotScores(filtered, image, 0xFFFF8800)
// Write the image to this coanvas
const canvas = DOM.canvas(w, h);
canvas.getContext("2d").putImageData(image, 0, 0)
return canvas;
}
marlon_and_ben = {
let winner, wscore, lscore, year, result = [];
// Games where Marlon won
const marlon_v_ben = col('winnername').eq('Marlon Hill').and(col('losername').eq('Ben Schoenbrun'));
// Games where Ben won
const ben_v_marlon = col('losername').eq('Marlon Hill').and(col('winnername').eq('Ben Schoenbrun'));
scrabble
.filter(marlon_v_ben.or(ben_v_marlon))
.scan((idx) => {
result.push({'winner': winner(idx), 'wscore': wscore(idx), 'lscore': lscore(idx), 'year': year(idx)});
}, (batch) => {
winner = col('winnername').bind(batch);
wscore = col('winnerscore').bind(batch);
lscore = col('loserscore').bind(batch);
year = col('year').bind(batch);
});
return result;
}
{
const game = scrabble.get(game_num);
return `${game.winnername} beat ${game.losername} ${game.winnerscore}-${game.loserscore} in ${game.year}`;
}
{
let win_total = 0, lose_total = 0;
for (let game of scrabble) {
win_total += game.winnerscore;
lose_total += game.loserscore;
}
return `Total Winner Score: ${win_total}, Total Loser Score: ${lose_total}`;
}
{
let total = 0;
for (let game of scrabble.filter(col('winnername').eq(player).or(col('losername').eq(player)))) {
total += game.winnername == player ? game.winnerscore : game.loserscore;
}
return `${player}'s Total Score: ${total}`;
}
{
return "Comment out this line to try it out. It will be slow!!"
let total = 0;
for (let game of scrabble) {
if (game.winnername == player) total += game.winnerscore;
else if (game.losername == player) total += game.loserscore;
}
return `${player}'s Total Score: ${total}`;
}
import {date, precipitation, temp_max, temp_min, wind} from "@mbostock/manipulating-flat-arrays"
seattle = arrow.Table.new(
// data
[
arrow.DateVector.from(date),
arrow.Float64Vector.from(precipitation),
arrow.Float64Vector.from(temp_max),
arrow.Float64Vector.from(temp_min),
arrow.Float64Vector.from(wind)
],
// names
["date","precipitation","temp_max","temp_min","wind"]
)
Array.from(
seattle.filter(arrow.predicate.col('precipitation').gt(50)),
row => row.date
)
One platform to build and deploy the best data apps
Experiment and prototype by building visualizations in live JavaScript notebooks. Collaborate with your team and decide which concepts to build out.
Use Observable Framework to build data apps locally. Use data loaders to build in any language or library, including Python, SQL, and R.
Seamlessly deploy to Observable. Test before you ship, use automatic deploy-on-commit, and ensure your projects are always up-to-date.