Sentiment plot of music albums

This readme only contains the technical details on how to run the project. If you are interested in the reason why I set up the project, visit http://www.edriessen.com/avicii/.

Analysing your favourite artist

Plotting the sentiment of the songs of one of your favourite artists takes two steps:

Analyse the song texts.
Plot the sentiment data.

Let's dig in.

1. Analyse the songtexts

To run an analysis, you'll need a connection to the Google Cloud Natural Language API. Set up a project in the Google Cloud console and add your credentials.json to the project root folder. After that, you'll need a .txt file for each song that you want to analyse. Format each song file this way: YYYY_albumname_songindex_songtitle.txt. For the song Wake Me Up by Avicii, this would be: 2013_true_1_wakemeup.txt. When you have the files ready, you can use analyse.py to run the sentiment analysis. You can call save_array_of_dicts_to_excel(). The function takes two arguments:

Array of dicts (the songs). Use the get_song_sentiment('songs/avicii') function to generate this list. Refer to the correct path location of your songs. In my example, the path is songs/avicii.
Name extension of Excel file. The function will save the file to the output/ folder. The file name will be song_sentiment_your_extension.xlsx.

A full example looks like this:

save_array_of_dicts_to_excel(get_song_sentiment('songs/avicii'), 'avicii_discography' )

2. Plot the sentiment data

You can visualise the data from the visualise.py file. You can call the scatter_plot_from_datafrome function. It takes three arguments:

Data frame. Use convert_xlsx_into_dataframe('output/song_sentiment_avicii_discography.xlsx') and refer to your file of choice.
Colour type. Use 'sentiment' for red, grey and green colours. Use 'album' for custom album colours (see paragraph below).
Magnitude magnification. The number of pixels a circle increases in size per magnitude point.

Here's a full example:

scatter_plot_from_dataframe(convert_xlsx_into_dataframe('output/song_sentiment_jj_first_three_discography.xlsx'), 'album', 28)

Using custom album colours.

When you set the colour type to 'album', you'll need to modify a part of the visualise.py file. Look for the part that says if color_filter=='album'. Modify the values inside the if statement to match your album names and colours of choice. Here's the example I used for Avicii's albums:

    if color_filter=='album':
        for value in dataframe['album']:
            if value == 'true':
                colors.append('#0E4691')
            elif value == 'stories':
                colors.append('#DC5463')
            else:
                colors.append('#EABE67')

Preparing a plot for print

Making your plot ready for print is easy. Simply save it as an .eps file. Open it in your editor of choice and remove all the unwanted elements (e.g. the graph elements and white background). After that, save it in your format of choice.

To do

The project is currently in a rough state. Things I'll be working on are:

Updating the project code to make it easier to use.
Make it easier to prepare a plot for print.