Grant Beasley

Interested in data, stats, coffee and music. Data Scientist at SaleCycle. Previously worked in pro rugby. All views are my own.

[Grant Beasley @grantbeasley]

I'm speaking at @Neo4j Meetup tomorrow evening (https://t.co/8IL6LnfGqh) about recommendations with Neo4j and Python. If you're around the area then come join the fun

[Grant Beasley @grantbeasley]

Continuing from last week's post - Clustering our @spotify data with @databricks, PySpark, @MLflow and Hyperopt. A little Brucey Bonus for anyone who manages to get all the way to the end. https://t.co/PtrRpvUz0N

[Grant Beasley @grantbeasley]

To carry on the fun of the Bank Holiday weekend, here's this week's blog post - Prog Rock, @databricks and @Spotify. Enjoy https://t.co/8C6Apeq09S

[Grant Beasley @grantbeasley]

This week's blog post - a look at causal inference and the dreaded omitted variable bias via the medium of rugby. Thanks to @MatheusFacure for the excellent teaching resources
https://t.co/OXr9HWtoxG

[Grant Beasley @grantbeasley]

This weeks blog post - using dbt Cloud (@getdbt) to create robust and scalable pro rugby analytics pipelines

https://t.co/gLQR7Yjit4

[Grant Beasley @grantbeasley]

This week's blog post - a smorgasbord of acronyms - an introduction to dbt @getdbt using the NCAA basketball dataset through GBQ (Google BigQuery) (not a real acronym but added for comic effect...😑) https://t.co/9NmQ09d9zw

[Grant Beasley @grantbeasley]

So here it is, the one we've all been waiting for - the fourth and final part of the @kaggle H&M recommendations challenge. Combining some @neo4j, Cypher queries, image embeddings and matrix factorization methods to create product recommendations https://t.co/5rxtIXNCcv

[Grant Beasley @grantbeasley]

Part 3 of the H&M recommendation challenge - collaborative filtering with implicit feedback. A little bit of @neo4j and a lot of code borrowed from @eprosenthal
https://t.co/tbPMPE4i3y

Check out his original post at https://t.co/QzfGVrDUjJ

[Grant Beasley @grantbeasley]

Part 2 of the H&M recommendations challenge - image embeddings and @neo4j Fabric. Creating recommendations from product images. So excited by this one I've shared it a day earlier than normal... https://t.co/yDZ7eOXUvT

[Grant Beasley @grantbeasley]

Part 1 of the @kaggle H&M recommendations blog series is now up - loading the data with @neo4j and a nice little bit of exploratory analysis to get us started. The excitement is palpable. https://t.co/qRZatzyqGr

[Grant Beasley @grantbeasley]

@CJLovesData1 @kaggle @neo4j Yeah that's no problem at all. Agree it's always good to see different approaches, I doubt mine is the most efficient way anyway 😂

[Grant Beasley @grantbeasley]

A sneak peak of this weeks blog post where we'll be looking at creating recommendations for the @kaggle H&M recommendations competition. Step 1 - exploring the data with @neo4j - blog post to follow in the next couple days...

[Grant Beasley @grantbeasley]

@CJLovesData1 @kaggle @neo4j This link has the info for how I loaded the graph and some EDA in python if that fits with the theme of your video?https://t.co/2XkFdUeiTp

[Grant Beasley @grantbeasley]

@CJLovesData1 @kaggle @neo4j Yeah sounds good, here's the initial script I created to ingest the data into the database https://t.co/4gBE1RiBZn. Working around memory issues with CALL and TRANSACTIONS

[Grant Beasley @grantbeasley]

One for the rugby/stats fans this week looking at the best way to model the number of tries scored in a match https://t.co/Sb7QyuM00H

[Grant Beasley @grantbeasley]

@brisvegas1 A brief look at graph vs hugging face document embeddings. May be of interested to you after your original question... https://t.co/b8uZiyE4yG

[Grant Beasley @grantbeasley]

This weeks blog post - Graph Embeddings and Named Entity Recognition via the medium of Sports Science with @neo4j and @JSCRonline - Part 1
https://t.co/iruu64dvtY

[Grant Beasley @grantbeasley]

And to conclude the mini series on Graph Embeddings and Named Entity recognition with @neo4j and @huggingface, here's part 3 - https://t.co/jCPtsvseLi - Graph vs Hugging Face embeddings

[Grant Beasley @grantbeasley]

Part 2 of last of the latest blog post - Centrality, Information flow and entity extraction with @neo4j - https://t.co/2Nkhqv0Llu

[Grant Beasley @grantbeasley]

@brisvegas1 @neo4j @JSCRonline Sounds interesting. Not something I've looked at previously but will try and look into it as part of the next few blog posts