#jupyter

[fray-ah]

april 12, 2021

accidentally cut my finger while cutting onions but otherwise it’s been a good day heh

[lizardposting]

opens jupyter and my fingers immediately start hurting

[mintlimeginger]

Playing around with the corpus of all transcripts from all episodes of CR. After removing the most common English words, these are the top 4-word groupings. Mentally constructing my prototypical CR episode now.

[greenglasstudy]

finally got a monitor so I can code easier! Goodbye squinting 😂 Right now I am working on a code for a robust quadratic solver. I’m also working on using spyder and latex for documentation. Really fun stuff!

[junolockscreens]

Milky Way. Photo by Scott Lacasse, Lacasse Photography, 2016.

[jadagul]

While I'm complaining about Jupyter: one thing it lacks (as far as I know) that Mathematica et al. have, is the ability to launch a notebook without tying it to a file.

When I was running Mathematica, I would often use it as basically an overpowered calculator. And this worked great, because I'd just open a window, run whatever calculations I wanted, and then close it when I was done. If nothing you're doing ever spans more than a single cell, and you don't save any of it, there are no problems here.

With Jupyter and Sage, I can in principle do the same thing. Except from what I can tell, jupyter-notebook will always open a file browser, and then you have to open a specific file to run commands—even if you have no desire to save any of them.

(And this interacts a lot with the discussion here. Jupyter is a fine interface for actually using as a temporary calculator, except that it seems to fight you on the possibility of doing that. What I really want is an interpreter shell that plays nice with the GUI. Once you start trying to save your work and build something with complicated architecture, these problems start biting you.

[epi-vet]

Good Omens Playlists top 100 songs, a playlist by epivet on Spotify

Top 100 songs based on ~1300 Good Omens-themed playlists

In my Introduction, I mentioned my Good Omens playlist project. My goal in this post it to explain my inspiration, methods, and next steps. For anyone only interested in the music, the URL will take you to a Spotify list of the top 100 songs on Good Omens- themes playlists. I have several other related public playlists if you want to click on my Spotify account (epivet).

Inspiration: Like many others, I directed my love of Good Omens into a playlist soon after watching it. While building that playlist, I searched Spotify for other Good Omens playlists for inspiration. After discovering the sheer volume of playlists, I wanted to quantify one way the Ineffable fandom expresses itself. Hence, my #quantified fandom tag. I started with ~350 playlists in September. Thanks to the ever-growing number of Good Omens-themed playlists and some improved search strings, I downloaded ~1400 playlists in December.

Methods: I identified Good Omens-themes playlists in Spotify by searching a set of terms and examining all results under playlists. Search terms included: Good Omens, Ineffable, Aziraphale, sauntered, and “nice and accurate”. I performed a quick review of each playlist before marking it for inclusion. Some users had more than one Good Omens playlist. Playlists were downloaded using Exportify (https://github.com/watsonbox/exportify) and then stitched into a single CSV. Playlist name is not an attribute automatically included, so userid was used to link back to playlist names and later to playlist classifications (Ineffable Husbands-themed, Crowley-themed, etc.).

I removed a few extremely outlists lists (>750 songs) for manageability, resulting in 78,144 rows of data (where each row represents a song on a specific playlist). Automated and manual formatting corrections addressed song title inconsistences, ensuring, for example, that “Good Old Fashioned Lover Boy” and “Good Old Fashioned Lover Boy – 2011 Remaster” were treated as the same song for the purposes of this analysis. I moved this dataset to a Jupyter notebook for analysis. I calculated Song Title frequency for all data and for specific artists using value_counts(). Similar analyses were performed for subsets of the full dataset (e.g. only playlists with the title “Ineffable Husbands”.) The top 50 songs on Ineffable Husbands playlists is also available on my Spotify.

Each song has a timestamp of when it was added to the user’s playlist. My next data dive will be to look at the spread of popular songs over time. I expect many Queen songs will start high with limited room for increased popularity. I am particularly interested in finding songs that were uncommon for months and then surged in popularity. Since I downloaded the playlists in December, roughly half a year after the debut of Good Omens, I do not have playlists that were created after my September pull and then deleted before my December pull. The playlists I have in both datasets may be an interesting subset for exploration. I am curious if playlists in the fandom start to move toward a common canon slowly over time. I know anecdotally that many Ineffable fans, myself included, discovered Hozier solely through the ubiquity of From Eden on playlists.

Next steps: I am also working on code that will let me tell people how common the songs on their own playlists are. Feel free to send questions or requests! I don’t have a ton of free time, but I definitely am using this project as an excuse to expand my Python and occasionally even Excel skills.

[iamprogrammerz]

How to work with Jupyter Notebook in Visual Studio Code ☞ http://bit.ly/39OgMv8 #Jupyter #VSCode