With the holiday and all, my schedule over the last few days has been very different and my days spent in alternate pursuits. I hope that you all had a wonderful Thanksgiving with your family of choice, if that is your thing.
Gina and I have been doing Thanksgiving on our own for many years now and we enjoy the downtime to relax, watch some movies, and eat yummy food. For many years she used to make cornish hens but starting last year she has been making duck breast and they came out perfect. We are having the leftovers tonight!
I have mentioned in the past that Gina is a goddamn Superhero and she proved it once again by launching her new program: From Silence to Self: A Guided Story Mining Journey – and I couldn’t be more proud of her. She is continuing her great work focusing on voice and storytelling to help others who are on the same journey.
I took the time to work on a little project of my own. I have talked about the last.fm site in the past, but essentially it has tracked almost everything I have listened to on my computer for almost 20 years now, about 150,000 plays. Over that time period, the data has gotten pretty messy.
Last.fm is great but the different music services (and before that MP3s) name things differently, and last.fm treats all the variations of an album or song as completely different, so, for example, I have plays from Frampton Comes Alive! and Frampton Comes Alive! (25th Anniversary Edition) and Frampton Comes Alive! (35th Anniversary Deluxe Edition) – and if it doesn’t have the ! after Alive, that is a whole different album as well. Even at the artist level, The Allman Brothers Band is different than Allman Brothers Band.
It’s a mess and even though you can go in and edit to consolidate everything, it’s hard to find the issues to correct them unless you happen to be browsing your library and notice the inconsistency.
Enter the last.fm Scrobble Analyzer – a program that scans your library and identifies potential issues in your scrobbles (that is what they call each logged play) so that you can fix them.
The developer community around last.fm is great and there are so many tools that help you scrobble your listens, or visualize your statistics in cool ways, but I have never come across something that helped identify these inevitable data errors.
So I built it.
It’s a very early version, so for now you you have download an export of your library file (I recommend using lastfmstats.com for this) but eventually it will access it directly via the API. Even though it’s still early in development I am finding it insanely useful so I wanted to get it out to people now so that they can start using it and help test and provide feedback on future versions. You can check out the github Repository for more details.
Here are some screenshots of the program in action:
This is the issue list – it shows the types of issues it finds and the number of each issue.
This is an example of a specific issue, songs by an artist that do not have an album associated with it. In this case, I have 110 different tracks by Gov’t Mule that don’t have an album, accounting for 348 total plays. That can really skew my charts!!
This project is a collaboration between me and the AI Claude. I wrote the specifications, defined the problem and all the functionality that I wanted, and it wrote the code – it’s all a single file and it imports all 150,000 lines in just a few seconds. Amazing.
I never let AI write for me but as a research tool (that I verify) and to summarize data and to write code, it is invaluable. I have had the vision of what I wanted this to do for years but without the skills to build it. Now it is built.
I have big plans for future updates as well so this is an ongoing thing I want to keep working on. I shared it on Reddit in the last.fm group and the response has been overwhelmingly positive which feels really good. Again, I have been using the tools that others have developed for years so it feels good to give back to the community.
Now that things are returning to normal (at least for a few weeks) you can expect a more regular posting of the random songs and such.
Your thoughts?