Matching users using Node.js on my website

Question

I am currently developing a "dating" website as a school project. I'm using Node.js, Express, my SQL, and Pug (Jade).

So far, my users have profile pages and in database I have the following information regarding them:

• Age
• Sexual orientation
• Sex
• Location (latitude and longitude, country, and exact city)
• Tags (words that defines them the best)

Now that I have all of this, my user has to be able the search by:

• Age
• Location and tags.

I am supposed to render by order, and the first match should always be the closest by location.

How do I sort all of theses information to check if any of my users can match one or several persons?

Answer 1

You could loop through all users in your database and calculate a score. Points are e.g. rewarded based on how near they are. In the end, the person with the most points will be the best match.

---

Here is an example with some randomly generated data.
I made the assumption

• that every person is interested in the opposite sex
• that latitude and longitude are coordinates on a plane in order to use the Pythagorean theorem to calculate distance

let data = [
  {
    'name': 'John Doe',
    'sex': 'male',
    'pos': [ 43.036871, -89.324967 ],
    'tags': [ 'general', 'basic' ]
  },
  {
    'name': 'Amy Schmidt',
    'sex': 'female',
    'pos': [ 39.48586, -121.387316 ],
    'tags': [ 'honest', 'uneven' ]
  },
  {
    'name': 'Robert Summers',
    'sex': 'male',
    'pos': [ 33.657366, -86.643871 ],
    'tags': [ 'efficient', 'psychotic' ]
  },
  {
    'name': 'Steven Walls',
    'sex': 'male',
    'pos': [ 43.484856, -83.849829 ],
    'tags': [ 'huge', 'grumpy' ]
  },
  {
    'name': 'Elizabeth Bateman',
    'sex': 'female',
    'pos': [ 38.886231, -99.306865 ],
    'tags': [ 'heavy', 'goofy' ]
  },
  {
    'name': 'Robert Galusha',
    'sex': 'male',
    'pos': [ 29.713645, -95.534338 ],
    'tags': [ 'vast', 'depressed' ]
  }
];

function search(person, tags) {
  let scores = { }, distances = { },
      lat = person.pos[0], lng = person.pos[1];
  
  data.filter(user => user.name !== person.name).forEach(user => {
    scores[user.name] = 0;
    scores[user.name] += user.tags.filter(tag => tags.indexOf(tag) > -1).length;
    scores[user.name] += user.sex !== person.sex ? 1 : 0;
    
    let dlat = Math.abs(lat - user.pos[0]), dlng = Math.abs(lng - user.pos[1]),
        distance = Math.sqrt(Math.pow(dlat, 2) + Math.pow(dlng, 2));
    
    distances[user.name] = distance;
  });
  
  // Needed to normalize the distances
  let maxDistance = Object.values(distances).sort((a, b) => b - a).shift();
  
  for(let name in distances)
    // Substract the normalized distance from 1, so: shorter distance = more points
    scores[name] += (1 - distances[name] / maxDistance);
  
  // Sort by score; the best match is the first element
  return Object.entries(scores).sort((a, b) => b[1] - a[1]);
}

console.log(search(data[0], [ 'honest', 'vast' ]))

If you want some factors to have a heavier impact on the overall score than others, you can multiply them by a certain weight.