What's next?

1. Put yourself in context

After a hard week at IronHack...

Relax

This club where they play the music that you like

Read a book

Go for sports

Go party

2. The reality

The session is not customised for that night's public

You want to go to a club where they play the music that you like

The DJ maybe has a session prepared for the general public

3. Solution

Build a music recommender that allows the DJ to adapt the session to the public

A recommender that takes the songs that the public requests, during the night, and gives back a song adding it to session

HOW?

4. Target sectors

- DJ's

 

- Clubs/disco

 

- Bars/Pubs

 

 

 

 

5. Music Recommender: Workflow

User's Song

Hot 100

Hot 100 DB song

Spotify DB song

Get audio features

Classify according to model (All songs)

Choose a song

Shows possible songs

Confirm artist

Spotify

Ask artist

Ask artist

DJ's playlist

DJ's playlist

6. DEMO TIME!

7. Limitations from this version

- Size of the database

 

- Missing data while getting information from Spotify

 

- Less accurate classification using just audio features and the amount of tracks in the DB or cluster method. 

 

- Possible repeated songs as recommendations

 

8. Future improvements

• Add the recommended songs automatically to the playlist/session that the Dj is playing  

 

• Expand the database:                                              

 

 

 

• Use other kind of Machine Learning technics.     

 ​

• Add an evaluation from the DJ's on the effect of the song recommended

• Improve the model and make more accurate recommendations
• Have more options in terms of  recommendations

Thank you for your time