z = require('https://bundle.run/zebras@0.0.11')
//get more information from an observable notebook: https://observablehq.com/@nickslevine/introduction-to-zebras-a-data-analysis-library-for-javascr
//get how to use all the functions: https://zebrasjs.com/Z.html#.sortByCol
rainAmounts = [
18.3,
19,
14.1
]
rainDates = Array.from({length: 3}, (_, i) => Date.now() - 1000 * 60 * 60 * 24 * i)
//three elements, two arrays; loop to get through the
newRainfall={
let rainArray=[]
for (var i=0; i<rainAmounts.length; i++){
let currRow={rainAmount: rainAmounts[i],
date: rainDates[i]}
//now we have an object, this is the way zebras considers as a dataframe
rainArray.push(currRow)}
//storing every element of currRow into rainArray's end
return rainArray
//remember: type in "rai" then "tab", you can get access to things
}
rainfall = {
let rainDF = []
for (var i=0; i<rainAmounts.length; i++) {
rainDF.push({precipitation: rainAmounts[i], date: rainDates[i]})
}
return rainDF
}
//object, like the observations; arrays like the columns
daySeries = ["Monday", "Tuesday", "Wednesday"]
myNewDF = z.addCol("day", daySeries, rainfall)
//add a column called "day" holding the values of daySeries in the dataframe rainfall
//now add a column to all objects in the df
columns = Object.keys(rainfall[0])
//the key of object is the name of columns
shape = ({rows: rainfall.length, columns:Object.keys(rainfall[0]).length})
//rainfall是一个存了很多object的df，所以rainfall.length即返回观测数；column的数量需要根据key的数量判断（可以将key视为variable)；注意每个key必须根据一个object进行提取，而不是对于整个df。
migrationUrl = "https://gist.githubusercontent.com/cesandoval/b834ac93c07e03ec5205843b97f68017/raw/90951b60444376eebfcbbee9beca00c083f489f6/API_SM.POP.NETM_DS2_en_csv_v2_10473747.csv"
//input csv file from github
migration = d3.csv(migrationUrl)
// knowledge about d3: https://github.com/d3/d3
z.head(5, migration)
migration.length
({rows: migration.length, columns:Object.keys(migration[0]).length})
//way to get rows and columns
migrationColumns = Object.keys(migration[0])
firstRow = Object.values(migration[0])
lastRow = Object.values(migration[migration.length-1])
randomRow = Object.values(migration[Math.floor(Math.random() * migration.length)])
numericalColumns = migrationColumns.slice(0, migrationColumns.length-4)
//use "slice" to get everything except the last 4; get the subset from 0 to the last-4.
migrationParsed = z.parseNums(numericalColumns, migration)
//turn the values into numbers using "z.parseNums"
z.head(5, migrationParsed)
//z.head is determining the number of rows displayed
countryName = z.getCol('Country Name', migrationParsed)
// 10th row of our Dataframe:
countryName[9]
z.describe(z.getCol('1962', migrationParsed))
//summary statistics of that column
z.valueCounts(z.getCol('Indicator Code', migrationParsed))
//"z.valueCounts" returns the fequency of each value within 1 column; i.e. return the number of unique values
z.unique(z.getCol('Indicator Code', migrationParsed))
//give an array of unique values
z.filter(r => r['1962'] > 0, migrationParsed)
migrationParsed
//the previous calculation will not change the original df
function(row){
return row['1962']>500
}
//similarly, this will work the same way.
z.filter(r => r['Country Name'] == "China", migrationParsed)
//condition + df to work with
z.filter(r => r['1962'] > 0 && r['2012'] < 0, migrationParsed)
//We can use argument 'asc' to reverse the sorting
migrationSorted = z.sortByCol('2017', 'des', migrationParsed)
//sort by aa given column; des: larger to smaller
z.filter(r => r['1962'] > r['2012'] , migrationParsed)
z.head(5, z.pickCols(['Country Name', '2017'], migrationSorted.slice(1)))
z.filter(r => r['Country Name'] == "Cayman Islands", migrationParsed)
isNaN(z.getCol('1960', migrationParsed)[1])
z.filter(r => Number.isNaN(r['1962']) == false, migrationParsed)
//get those rows with no NaN for year 1962
z.getCol('1960', migrationParsed)
z.dropCol(["1960"], migrationParsed)
migrationMetaUrl = "https://gist.githubusercontent.com/cesandoval/b834ac93c07e03ec5205843b97f68017/raw/90951b60444376eebfcbbee9beca00c083f489f6/Metadata_Country_API_SM.POP.NETM_DS2_en_csv_v2_10473747.csv"
migrationMeta = d3.csv(migrationMetaUrl)
//d3.csv - get a comma-separated values (CSV) file.
migrationMeta.length
migrationMetaColumns = Object.keys(migrationMeta[0])
z.head(5, migrationMeta)
md`
> \`\`\`javascript
(static) merge(dfLeft, dfRight, leftOn, rightOn, leftSuffix, rightSuffix) → {df}
\`\`\`
`
//the join is on the left, thus right/ left order matters
migration_join = z.merge(migrationParsed, migrationMeta, "Country Code", "Country Code", "_df1", "_df2")
// migration_join = migrationParsed.map(x => Object.assign(x, migrationMeta.find(y => y['Country Code'] == x['Country Code'])));
migration_2017 = z.getCol('2017', migration_join)
z.print([
{
minMigration: z.min(migration_2017).toFixed(5),
maxMigration: z.max(migration_2017).toFixed(5),
meanMigration: z.mean(migration_2017).toFixed(5),
medianMigration: z.median(migration_2017).toFixed(5),
stdMigration: z.std(migration_2017).toFixed(5)
}
])
migration_2017_Piped = z.pipe(
[
z.parseNums(['1992','1997','2002','2007', '2012', '2017']),
z.getCol('2017'),
z.mean()
]
)(migration_join)
//chaining law. apply the functions one by one, with no need to passing a new df name every operation you do.
migrationByRegion = z.groupBy(x=>x.Region_df1, migration_join)
migrationCountRegion = z.gbCount("Region", migrationByRegion)
d3 = require("d3")
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.