Houses, Arrays and Higher Order Functions

In the previous post we made use of higher order functions to generalise the summation function. Using HOFs we managed to implement a solution that was much simpler - as defined by Rich Hickey over here - than its iterative (and less evolved) cousin. In this post we are going to apply the same technique on arrays and apply this on a dataset of house prices. Who knows maybe this techniques will help you find your dream home!

Dataset

The dataset we are going to use throughout this example can be downloaded from here. This dataset contains the following fields [^1]:

MLS: Multiple listing service number for the house (unique ID).
Location: City/town where the house is located.
Price: The most recent listing price of the house (in dollars).
Bedrooms: Number of bedrooms.
Bathrooms: Number of bathrooms.
Size: Size of the house in square feet.
Price/SQ.ft: Price of the house per square foot.
Status: Type of sale. Three types are represented in the dataset: Short Sale, Foreclosure and Regular.

Filtering the data

As a first exercise, let us find all houses in Cambria which have four bedrooms [^2]:

function housesInCambriaFilter(array) {
    var passed = [];
    for(var i=0; i < array.length; i++){
        var house = array[i];
        if (house.location === "Cambria" && house.bedrooms == 4) passed.push(house);
    }
    return passed;
}
housesInCambriaFilter(REALESTATE_DATA);

There is only one house that matches our criteria:

[ { mls: 147819,
    location: 'Cambria',
    price: 332000,
    bedrooms: 4,
    bathrooms: 4,
    size: 1872,
    'price-sq-ft': 177.35,
    status: 'Foreclosure' } ]

Now let us find all houses where the price per square foot is less than 20:

function lowPricePerSquareFoot(array) {
    var passed = [];
    for(var i=0; i < array.length; i++){
        var house = array[i];
        if (house['price-sq-ft'] < 20) passed.push(house);
    }
    return passed;
}
lowPricePerSquareFoot(REALESTATE_DATA);

There are two houses which match this criteria, interestingly they are both in Santa Maria-Orcutt

[ { mls: 148168,
    location: 'Santa Maria-Orcutt',
    price: 29000,
    bedrooms: 2,
    bathrooms: 2,
    size: 1500,
    'price-sq-ft': 19.33,
    status: 'Foreclosure' },
  { mls: 154462,
    location: ' Santa Maria-Orcutt',
    price: 26500,
    bedrooms: 2,
    bathrooms: 2,
    size: 1344,
    'price-sq-ft': 19.72,
    status: 'Regular' } ]

Observe that the two methods are almost identical except for the following filtering code snippets:

if (house.location === "Cambria" && house.bedrooms == 4) passed.push(house);
if (house['price-sq-ft'] < 20) passed.push(house);

Using higher order functions we can separate the filter from the iteration logic as follows:

function filter(array, test){
    var passed = [];
    for(var i = 0; i < array.length; i++){
        if (test(array[i])){
            passed.push(array[i]);
        }
    }
    return passed;
}
filter(REALESTATE_DATA, function(house){
    return house.location === "Cambria" && house.bedrooms == 4;
});

filter(REALESTATE_DATA, function(house){
    return house['price-sq-ft'] < 20;
});

This results in code which is much more simple and clean. In ECMA-262 5th edition (aka ES5) we have access to the filter function directly on arrays and thus we can write this code as follows:

REALESTATE_DATA.filter(function(house){
  return house.location === "Cambria" && house.bedrooms == 4;
});

REALESTATE_DATA.filter(function(house){
  return house['price-sq-ft'] < 20;
});

Wow now that is much simpler Using ECMAScript6 we can write the same piece of code using a more expressive closure syntax as follows:

filter(REALESTATE_DATA, 
    (house) => house.location === "Cambria" && house.bedrooms == 4)
filter(REALESTATE_DATA, 
    (house) => house['price-sq-ft'] < 20 );

Mapping Data

Now that we have an array of objects representing the houses produced by filtering the REALESTATE_DATA, we want to produce an array with just the MLS attribute. Let us do this on the Cambria filter we defined earlier:

var housesInCambriaFilterData = housesInCambriaFilter(REALESTATE_DATA);
function getMLS(array) {
    var transformed = [];
    for (var i = 0; i < array.length; i++) {
        var house = array[i];
        transformed.push(house.mls)
    }
    return transformed;
}
getMLS(housesInCambriaFilterData);
// -> [ 147819 ]

Let us repeat this exercise and retrieve just the Price/SQ.ft on the low price per square foot filter we defined earlier:

var lowPricePerSquareFootData = lowPricePerSquareFoot(REALESTATE_DATA);
function getPricePerSquareFoot(array) {
    var transformed = [];
    for (var i = 0; i < array.length; i++) {
        var house = array[i];
        transformed.push(house['price-sq-ft'])
    }
    return transformed;
}
getPricePerSquareFoot(lowPricePerSquareFootData);
// -> [ 19.33, 19.72 ]

Notice that the only difference between the two methods is:

transformed.push(house.mls)
transformed.push(house['price-sq-ft'])

By creating a map function we can rewrite these two methods as follows:

function map(array, transform){
    var mapped = [];
    for(var i = 0; i < array.length; i++){
        mapped.push(transform(array[i]));
    }
    return mapped;
}

map(housesInCambriaFilterData, function(house){
    return house.mls;
});

map(lowPricePerSquareFootData, function(house){
    return house['price-sq-ft'];
});

The map function has also been standardized in ECMA-262 5th edition (aka ES5) and we can in fact use the map functions directly on arrays as follows:

housesInCambriaFilterData.map(function(house){
    return house.mls;
});

lowPricePerSquareFootData.map(function(house){
    return house['price-sq-ft'];
});

Using ES6 we can write the code as follows:

housesInCambriaFilterData.map((house) => house.mls);
lowPricePerSquareFootData.map((house) => house['price-sq-ft']);

Composing Filter and Map

Higher Order Functions start to shine when we need to compose functions. In fact the above code can be written without the use of the intermediate variables housesInCambriaFilterData and lowPricePerSquareFootData:

REALESTATE_DATA.filter(function(house){
    return house.location === "Cambria" && house.bedrooms == 4;
}).map(function(house) {
    return house.mls;
});

REALESTATE_DATA.filter(function(house){
    return house['price-sq-ft'] < 20;
}).map(function(house) {
    return house['price-sq-ft']
});

In ES6 we have some more syntactic sugar which we can sprinkle on top:

REALESTATE_DATA.filter(
    (house) => house.location === "Cambria" && house.bedrooms == 4
).map((house) => house.mls);

REALESTATE_DATA.filter((house) => house['price-sq-ft'] < 20)
    .map((house) => house['price-sq-ft']);

Reduce

Another common operation is the computation of a single value from array elements. The higher order function that represents this pattern is called reduce or sometimes fold. reduce can be implemented as follows:

function reduce(array, combine){
    var current = array[0];
    for(var i =1; i < array.length; i++){
        current = combine(current, array[i]);
    }
    return current;
}

array is the array on which the reduction will performed and combine is the function that will be used to reduces two values. So how can we use reduce to say sum up the array [1, 2, 3, 4]?

reduce([1, 2, 3, 4], function(a, b){
    return a + b;
});

Intuitively, we can think of reduce a function which inserting the combine operator between elements values:

(((1 + 2) + 3) + 4)
combine(combine(combine(1,2),3),4)

As with filter and map there is standard reduce function which we can use. The example above can be written as follows:

[1, 2, 3, 4].reduce(function(a, b){ return a + b});

In ES6 we can make use of the expressive closure syntax and write:

[1, 2, 3, 4].reduce((a, b) => a + b);

Let’s us now apply reduce on the dataset to find the average number of bathrooms in Cambria

housesInCambriaFilterData.map(function(house){
        return house.bathrooms;
    }).reduce(function(current, bathroom) {
        return current + bathroom;
    }) / housesInCambriaFilterData.length
    
// -> 4  (Remember there was only one data point!)

Finally, let’s find out the average listing price in our dataset

REALESTATE_DATA.map(function(house){
        return house.size;
    }).reduce(function(current, size) {
        return current + size;
    }) / REALESTATE_DATA.length 
// -> 1755.0588988476313       

Can you find the average price in each location?

Consider writing a higher order function group which first groups the data using a given criteria e.g. (house) => house.location. Best implementation gets a virtual chocolate :bowtie: .

This dataset has been adapted from here
REALESTATE_DATA contains the whole dataset.
Icon from Dave Gamez dribbble account. Great work dude!

Written on January 16, 2016