Scott Stoltzman

Definitely avoid using renv inside your project, computer immediately overheats

And look at those great front-end chops. Beautiful Responsive design #css

Before I did any filtering, the biggest thing I noticed was that the most popular (not inappropriate tags) were always #art #photo / #photography

Thought that was pretty cool to see.

I have to say, the #bluesky API is really fun to play with. Websockets are definitely less intimidating than you think. 3 files to make this page below #javascript #html #css

Track whatever specific words you want, it tallies up the hashtags in the posts.

tinyurl.com/hashtagexample

Would definitely use something like this

@econmaett.bsky.social - Any desire for read access to the hashtags located on Azure to do some #dataviz or #rtstats? DM for temporary read-only access. #duckdb integrates seamlessly. Data in parquet:
did (user ID), cid (post ID), hashtag, created_at, text

ALL hashtag posts being stored.

python-dotenv

seriously... someone needs to give up dotenv

Added some extra spam filtering and filters to help keep the data more accurate.

Looks like the bot isn’t posting to the same thread, nor is it resetting. Haven’t looked since Christmas. Looks like these hashtags are stabilizing in the top 3, same with combos. Will need to modify the script and swap it to 12 hours instead of 6.

Each tag (yellow) should grow relationships (USED_TAG) to users (green). #sql should link this account.

I'll be working on the visualization over the holidays to embed network diagrams for the neo4j implementation of @stoltzmaniacbot.bsky.social programming hashtag bot

Dang it, thought that it wouldn’t “tag” unless it turned it blue, will cut this down to once per day until I figure out how to make it stop. Also, #rstats made an overnight comeback! Scott Bot will be modified tonight, feel free to suggest anything

Happy the hashtag bot doesn’t actually hashtag. Table needs prettying up. I was convinced the R community dominated the programming hashtag world. But maybe it’s just the weekend/holiday slump

Whoa, sorry for the bot testing spam! Gonna ruin another hashtag for testing next time

#javascript only. 4 life

Looks like #python has the best rating for mutual posting with mentions tagged. #golang coming in last
Total nodes: 129k

#python = 19%
#rstats = 10%
#golang = 4%

Sampling method is questionable:)

Also, run fact @mackaszechno.bsky.social , you were in my recommendations list!

Yeah, took me a while to pick. It seemed as if the OSS + examples + documentation lowered the barrier to entry. First time using actual graph database rather than the in memory packages reading from CSVs

I’d love to share easy access to this #neo4j database for querying. Any ideas of the simplest way to do so without simply handing out user/pass?

Scraped #python and #rstats from the latest package I built to take bsky data directly into #neo4j to make recommendations for follows directly out of a cypher query. Impressed by the speed! Going to have to start following some new people! I live in a small bubble in the graph visualization.

Made a couple of updates to #bskydata to simplify imports. Also, considering adding a "write to network" function. What is the consensus on the best network visualization library? Here is pyvis - ChatGPT defaults, didn't do any customization.

pip install bskydata

#python #pyvis

Namespace related stuff always makes for a good R-ant.

R 4.0+ has made some improvements. hoping that modules get introduced somehow (seems impossible) rather than having to continue making new names for every single task

Don’t people realize that there’s a reason “Jupyter” rather than “Stataturn” notebooks became major tools?

#rstats #stata #python #julia

Once you see it, you can’t unsee it. Very cool!

BREAKING: #python bskydata now has parser functionality so that data can be consistent, clean, and customized. No more full JSON dumps unless you need it. This will save a lot of storage space.

pip install bskydata

followers.json, follows.json, search_terms.json shown.

BREAKING: #python bskydata now scrapes and writes data directly to #mongodb -- I need some help on how it should insert, but it does connect and insert records! Simply switch out one line, your DataWriter is now MongoDBDataWriter

pip install bskydata[mongodb]

Also, @brianokken.bsky.social - guessing you may have some knowledge about what the best testing resources are. I want to write tests but since it's mostly just moving data around from APIs, it feels incredibly difficult to understand if things are working properly. Any mocking strategy you like?

BREAKING: #python BlueSky data scraping with auto-pagination now writes directly to the cloud! Integrations with #azure #aws #gcp have been completed. Simply swap out AzureDataWriter, GCPDataWriter, S3DataWriter in your scraper.

pip install [azure, aws, gcp]

Examples:
github.com/stoltzmaniac...

BREAKING: BlueSky data scraping is now easier. New update to #python bskydata package allowing you to search and store #bluesky posts. Auto-pagination & automated local file storage. Built for simplicity.

pip install bskydata

github.com/stoltzmaniac...